Skip to main content

Full text of "Journal of the Society of Motion Picture Engineers"

See other formats

From the collection of the 

^^^^v ^^^0 

o Prelinger 
v Jjibrary 

San Francisco, California 

Journal of the 

Society of Motion Picture Engineers 



Report of the President LOREN L. RYDER 3 

Film-Collection Program HOWARD LAMARR WALLS 5 

The Motion Picture Theater JAMES FRANK, JR. 9 

Effect of Television on Motion Picture Attendance 


Navy Photography in the Antarctic. .CHARLES CURTIS SHIRLEY 19 
Motion Picture Photography at Ten Million Frames Per Second 


Comparison of Lead-Sulfide Photoconductive Cells with Photo- 

Volume Compressors for Sound Recording W. K. GRIMWOOD 49 

Some Distinctive Properties of Magnetic-Recording Media. . . . 


Wide-Track Optics for Variable- Area Recorders 

Trend Control in Variable- Area Processing .... F. P. HERRNFELD 97 

Sixty-Fourth Semiannual Convention 103 

Section Meetings 109 

Meetings of Other Societies Ill 

Book Reviews: 

"Informational Film Year Book, 1948" 
Published (1948) by the Albyn Press 

Reviewed by Lloyd E. Varden 112 

"The High-Current Carbon Arc," by Wolfgang Finkelnburg 
Published (1947) by the Office of Military Government for 

Reviewed by F. T. Bowditch 112 

Current Literature 114 

New Products 115 


Chairman Editor Chairman 

Board of Editors Papers Committee 

Subscription to nonmembers, $10.00 per annum; to members, $6.25 per annum, included in 
their annual membership dues; single copies, $1.25. Order from the Society's General Office. 
A discount of ten per cent is allowed to accredited agencies on orders for subscriptions and 
single copies. Published monthly at Easton, Pa., by the Society of Motion Picture Engineers, 
Inc. Publication Office, 20th & Northampton Sts., Easton, Pa. General and Editorial Office, 
342 Madison Ave., New York 17, N. Y. Entered as second-class matter January 15, 1930, 
at the Post Office at Easton, Pa., under the Act of March 3, 1879. 

Copyright, 1949, by the Society of Motion Picture Engineers, Inc. Permission to republish 
material from the JOURNAL must be obtained in writing from the General Office of the Society. 
Copyright under International Copyright Convention and Pan-American Convention. The 
Society is not responsible for statements of authors or contributors. 

Society of 

Motion Picture Engineers 

342 MADISON AVENUE NEW YORK 17, N. Y TEL. Mu 2-2185 



Earl I. Sponable 
460 W. 54 St. 
New York 19, N. Y. 

Clyde R. Keith 
120 Broadway 
New York 5, N. Y. 

Loren L. Ryder 
5451 Marathon St. 
Hollywood 38, Calif. 

Peter Mole 

941 N. Sycamore Ave. 
Hollywood 38, Calif. 




John A. Maurer Ralph B. Austrian 

37-01 31 St. 25 W. 54 St. 

Long Island City 1, N. Y. New York 19, N. Y. 

Robert M. Corbin 
343 State St. 
Rochester 4, N. Y. 

William C. Kunzmann 
Box 6087 
Cleveland 1, Ohio 


David B. Joy 
30 E. 42 St. 
New York 17, N. Y. 


Alan W. Cook 
25 Thorpe St. 
Binghamton, N. Y. 

Lloyd T. Goldsmith 
Warner Brothers 
Burbank, Calif. 

Paul J. Larsen 
508 S. Tulane St. 
Albuquerque, N. M. 

Gordon E. Sawyer 
857 N. Martel Ave. 
Hollywood 46, Calif. 



James Frank, Jr. 
426 Luckie St., N. W. 
Atlanta, Ga. 

William H. Rivers 
342 Madison Ave. 
New York 17, N. Y. 

Sidney P. Solow 
959 Seward St. 
Hollywood 38, Calif. 

R. T. Van Niman 
4431 W. Lake St. 
Chicago 24, 111. 


Herbert Barnett 
Manville Lane 
Pleasantville, N Y. 

Fred T. Bowditch 
Box 6087 
Cleveland 1, Ohio 

Kenneth F. Morgan 
6601 Romaine St. 
Los Angeles 38, Calif. 

Norwood L. Simmons, Jr. 
6706 Santa Monica Blvd. 
Hollywood 38, Calif. 

Report of the President 

rplms is THE fourth semiannual report which I, as President, have 
J_ had the honor of presenting to the Society of Motion Picture 
Engineers. It will be a review of the high lights of activity during 
the two-year period of my presidency. 

The scope of Society activity has been denned to include all phases 
of pictorial rendition of action; whether it be from film, as in motion 
pictures, electronics, as in television, or other device. As will be ap- 
parent later in this report your Society is doing an able job in each 
of these fields. 

Under the guidance of Paul J. Larsen the Society of Motion Picture 
Engineers was the initiating and only active body in gaining frequency 
allocations for theater television. This activity has been carried on 
through the years until now our efforts have gained recognition and 
individual companies in the industry are applying for and gaining fre- 
quency allocations which this Society made available to this industry. 

The Society of Motion Picture Engineers stands out as a world 
leader in thinking and action in the field of theater television. The 
"firsts" which were brought about as a result of Society demonstra- 
tions are numerous and interesting. Some of them are listed below: 

1 . The use of television facilities as a visual public-address system. 

2. The use of television facilities as a means of instruction both 
to show close-ups of equipment and instructor. 

3. The showing of members of the audience on the screen to the 
entire audience. 

4. The filming, fast processing, and television projection of per- 
sons entering a meeting room. 

5. The presentation of television on a twenty-foot screen. 

6. The presentation of a television-broadcast pickup of a sports 
event and the immediate playing of a film transcription of the same 
event in order that the audience might again watch the high lights of 
the sport in action. 

The papers on the art of television which are now published in the 
JOUKNAL stand out as the one authentic record and statement of 
technical fact in regard to all phases of television as applied to motion 

The standardization work of the Society holds an important posi- 
tion, as there are now more standards as applied to motion pictures 
than to any other industry. This is important to this industry for it 
* Presented October 26, 1948, at the SMPE Convention in Washington. 



is through standardization that we gain the universal and world-wide 
market which we enjoy for our product. 

The Society and the newly constituted Motion Picture Research 
Council have established definite co-operative procedures of handling 
technical problems of mutual interest. As of this writing, joint com- 
mittees exist in the fields of sound, standards, and test films. There is 
now no overlap between the test-reel activity of the Society and that 
of the Research Council. The efforts of one organization augment 
and supplement without conflict that of the other, which is very 

The Society of Motion Picture Engineers having outgrown its pre- 
vious quarters has now moved into a new suite of rooms in the Cana- 
dian Pacific Building, 342 Madison Avenue, New York 17. The en- 
trance-room number is 912 and the telephone number is Murray 
Hill 2-2185. The staff of the Society includes Boyce Nemec, Execu- 
tive Secretary, William H. Deacy, Jr., Staff Engineer, Helen M. 
Stote, JOURNAL Editor, Sigmund ]\1. Muskat, Office Manager, and 
secretarial help, Helen Goodwyn, Dorothy Johnson, Thelma Klinow, 
Ethel Lewis, and Beatrice Melican. 

We are proud of this staff and the work that they are doing. We 
want you to know that you can call upon them and your Society when- 
ever the Society or our staff may assist you. 

The JOURNAL of the Society has taken on a new format and a policy 
of giving to our readers in so far as possible the knowledge and infor- 
mation they are interested in reading. Our editorial staff has done an 
excellent job, and I am happy that the journals are now being issued 
on schedule which was not always possible during the war years. 

Mr. John A. Maurer is continuing to do a colossal job as Engineer- 
ing Vice-President. Both through committee and by personal effort 
his activity and accomplishments must -have reached all of you. 

The convention in Chicago emphasizing 16-mm and 8-mm motion 
pictures, the convention in New York with a theater engineering 
conference and exhibit, and the Hollywood convention with emphasis 
on motion picture production in television have all been most success- 
ful and a record of which we are justly proud. I know that the Wash- 
ington convention will be remembered along with these other suc- 
cesses of our Society. 

I want to thank you all for the support which I have received. It 
has been an honor and a pleasure to serve the Society, and as I change 
my status from that of President to that of Past-President, I shall 
carry on with you to aid and serve you in whatever way I may. 

Respectfully submitted, 
LOREN L. RYDER, President 

Film- Collection Program* 



Summary The Library of Congress attempted to restore to the screen 
the first twenty years of motion picture achievements by the optical printing 
of photographic paper rolls submitted for copyright registration from 1897 to 
1917, but Congress refused to provide funds for the work. The Academy of 
Motion Picture Arts and Sciences, with the co-operation of The Library of 
Congress, now undertakes to do this work. 

As CURATOR OF MOTION PICTURES in The Library of Congress, it 
was my privilege to appear before your Society in New York 
City pn May 5, 1943, and to announce at that time the results of 
photographic tests that had been conducted with a view to restoring 
to the screen the first twenty years of motion picture achievements. 1 
Sometime during the latter half of 1942, we at the Library had come 
to the conclusion that positive paper rolls of motion pictures submit- 
ted for copyright registration between 1894 and 1917 might be trans- 
ferred to celluloid by some special photographic process. In following 
through with this thought, we approached John G. Bradley, then 
chief of the Division of Motion Pictures in the National Archives, for 
whatever assistance he could give us. Carl Louis Gregory, who was 
serving with the Archives at that time as a photographic engineer, was 
assigned to the problem. After a few tests at optical printing by 
reflected light, Mr. Gregory found that these opaque paper prints 
could successfully be copied. 2 With this discovery, it became evident 
that the foundations of our motion picture art-science, hitherto con- 
sidered to be irretrievably lost, could now be resurrected through one 
major project. 

It was the Library's intention to proceed at once with the copying 
of the nearly two and one half million feet of the paper films com- 
prising the collection, but the exigencies of the war defeated it. The 
entire facilities of the motion picture division had to be given up 
wholly to the servicing of the seized films of enemy aliens to the vari- 
ous war agencies in behalf of the Alien Property Custodian. In the 
meantime, the paper rolls were carefully housed in a specially adapted 

* Presented May 21, 1948, at the SMPE Convention in Santa Monica. 


6 WALLS January 

room pending the end of the war and a more opportune time to con- 
tinue with the plan of restoration. 

By the end of 1946, the Library was ready to resume its intention. 
It was estimated that the work would cover a period of approximately 
five years and would cost nearly a quarter of a million dollars. The 
method of procedure was set forth in the Library's budget for the fis- 
cal year June, 1947, to June, 1948, but the Senate Appropriations 
Committee, after considering the Library's needs in this connection, 
refused to provide funds for the work. Moreover, it denied the 
Library's wish to assume the responsibility of establishing a much- 
needed national film collection an activity in which the Library had 
been engaged, with the assent of previous Congresses, since 1942. The 
Committee's refusal was upheld in the votes of the House and Senate, 
and the Library's motion picture program was abruptly halted. The 
divisional staff, which had grown to thirteen persons, had to be sum- 
marily dismissed. 

As one who had acquired a conversant knowledge of the collection 
of paper prints and their almost incredible value, I naturally became 
quite concerned about the future of the material. In a move to pre- 
vent further deterioration of the paper rolls and to get them copied to 
celluloid for their various social uses, I appealed to the Librarian of 
Congress to lend them to the Academy of Motion Picture Arts and 
Sciences if the Academy Foundation would undertake the task of 
copying them. He agreed to this, and I came on to Hollywood to 
present the case. 

Both Jean Hersholt, president of the Academy, and Walter Wanger, 
chairman of the Academy Foundation, quickly expressed their interest 
in a cultural project of such importance. I am happy to be here today 
as curator of the Academy's Motion Picture Collection and to tell 
you that the beginnings of our screen history will be saved after all. 
I would like to pay a special public tribute to Mrs. Margaret Herrick, 
the executive secretary of the Academy, whose intense interest in the 
proposed project kept the idea alive and going during the few weeks 
that were necessary for the Academy's Board of Governors to arrive 
at the final and affirmative decision. 

The Academy Foundation is pleased to have this opportunity to 
demonstrate briefly to the Society the results of the first efforts at 
converting the paper prints to celluloid. To us, it is a technological 
achievement that borders on the miraculous. We think that perhaps 
you will agree. The films were made by Film Effects of Hollywood 


on an optical printer designed and built by Douglas Heidanus. I 
would like to tell you briefly about our plans, about what the Acad- 
emy Foundation hopes to accomplish with these films as they come 
off the optical printer. 

Most of you are familiar with the workings of the Academy of 
Motion Picture Arts and Sciences, which had its inception in 1927. 
The Academy has striven, and is working today, to advance the arts 
and sciences of motion pictures and to foster co-operation among the 
creative leadership of the motion picture industry for cultural, edu- 
cational, and technological progress. During the more than twenty 
years of its operation, it has recognized outstanding achievements by 
conferring annual Awards of Merit that serve as a constant incentive 
within the industry and focus wide public attention on the best in 
motion pictures. It has conducted co-operative technical research 
and stimulated the improvement of methods and equipment. 

For some time, the Academy has wanted to extend its operations to 
include a motion picture service along closely defined social and his- 
torical lines. It feels* that the motion picture medium should be pre- 
served in a collection of films so brought together as to provide an 
important national source of research material for students of the 
cinema and other types of scholarly users. It believes that this is the 
only way by which the history of the motion picture can be written 
without conjecture; it wants to make it possible for the story of the 
motion picture to be effectively assembled for visual presentation in 
the classrooms of universities, colleges, and high schools as a part of 
their regular curricula. 

The construction of such a collection to these ends is now under way 
with The Academy Foundation supporting the Academy's aim. The 
films that will be made from the photographic paper prints lent to us 
by The Library of Congress will, of course, serve as the foundations of 
the collection, while a carefully prepared plan of selection (designed to 
eliminate as far as possible mere personal judgment) will be put into 
operation to procure for preservation other important cinematic 
landmarks made up to current times. 

The freedom of selection and the consequent magnitude of the 
general collection will depend on the availability of funds that can be 
obtained to carry on the work as we progress. Contrary to common 
belief, the Academy is not supported financially by wealthy film- 
producing companies; its routine expenses are paid from the member- 
ship dues of cinematographers, actors, writers, and others. The 


Academy Foundation, now headed by Y. Frank Freeman, was estab- 
lished to receive tax-free donations and thereby support the broad 
cultural work of the Academy proper. It is through the support of 
this Foundation that the work on the paper prints will be accom- 
plished, that additional social and historical aspects of the motion pic- 
ture medium will be given expression. 

In concluding my remarks, I must add that the Foundation does 
not yet have the quarter of a million dollars required to convert the 
paper prints. We have only a shoestring and a lot of ambition. We 
must have donations if we are to save the art-science of the motion 
picture and put it to universal use. We are open to all the suggestions 
you care to make regarding sources that can help us do the job. 


(1) H. L. Walls, 'Motion picture incunabula in The Library of Congress," 
/. Soc. Mot. Pict. Eng., vol. 42, pp. 155-159; March, 1944. 

(2) C. L. Gregory, "Resurrection of early motion pictures," J . Soc. Mot. Pict. 
Eng., vol. 42, pp. 159-170; March, 1944. 

The Motion Picture Theater 


A new book carrying the above title and published by the Society of 
Motion Picture Engineers was announced in December, 1948. Motion 
picture exhibitors, purchasing agents, and theater architects will find a 
wealth of valuable information in its thirty-eight articles and 427 pages on 
the technical aspects of motion picture theater planning, construction, main- 
tenance, modernization, and theater television. 

Copies are available from the Society at $5.00, postage paid within the 
United States, except that copies sold in New York City are subject to an 
additional ten-cent city sales tax. Copies shipped outside the continental 
United States or its possessions are $5.50, postage paid. 

THE MATERIAL included in this book consists, for the most part, 
of papers presented and discussions which took place at the 
62nd Semiannual Convention of the Society of Motion Picture Engi- 
neers in New York City, October 20-24, 1947. 

Recognizing at that time that theater owners all over the United 
States were planning for the construction of new theaters and the 
modernization of existing ones, it was felt that the Society was both 
obligated and in a position to furnish these theater owners with the 
latest scientific information on major phases of theater design and 
construction. Furthermore, it was evident that architects assisting 
in these plans had not had an opportunity, except through the trade 
papers, both to learn of trends in other parts of the country and to 
exchange information with other architects regarding novel features. 

It was therefore decided early in 1947 that the major portion of 
the 62nd Semiannual Convention should be devoted to Theater 
Engineering. A preliminary survey conducted among theater organi- 
zations and circuits, and the enthusiastic response which it received, 
confirmed the Society's feeling that such a program was desirable. 
A special committee, consisting of leading theater architects and 
leading theater-circuit executives, responsible for construction and 
maintenance, was appointed to organize the conference. They deter- 
mined that individual sessions would be devoted to specific principal 
subjects of interest, and that experts, most of them representing 
manufacturers of theater equipment, would be invited to present 
short technical papers written in a style that would be of interest to 
theater owners, purchasing agents, construction men, and theater 
managers. They further planned that a major portion of each session 

10 FRANK January 

would be devoted to a discussion period permitting the various ex- 
perts to question one another and also to permit those attending the 
Conference to offer opinions or ask questions from the floor. In order 
to accomplish this, a specially designed public-address system was 
obtained so that all in the room could hear the entire discussion easily. 
To enhance the effectiveness of the Conference further, an exhibit 
of new and interesting theater equipment and materials was held in 
an adjoining room. The Society was successful in obtaining a num- 
ber of outstanding exhibits of this nature, which proved to be of tre- 
mendous interest to those attending the Conference. Credit for the 
success of the Conference, which made publication of this book pos- 
sible, is in no small part due to the Theater Conference Papers Com- 
mittee, as well as the Editorial Vice-President, Convention Vice- 
President, and the Staff of the Society. The members of the Special 
Papers Committee and the sessions for which they were responsible 
are as follows : 


Leonard Satz 
Century Circuit 

Auditorium Design Lighting 

Martin F. Bennett Wallace W. Lozier 

Radio Corporation of America National Carbon Company 

Floor Coverings Acoustics 

Charles Bachmqn Charles S. Perkins 

Warner Brothers Altec Service Corporation 

Television Safety and Maintenance 

Donald E. Hyndman Henry Anderson 

Eastman Kodak Company Paramount Pictures 

Ventilating and Air Conditioning; 
Television Projection Promotional Display 

Paul J. Larsen Seymour Seider 

Consultant Empee Construction Corporation 

To ensure the success of the Conference, great care was taken in 
the selection of Session Chairmen, all of whom had attained a high 
standing in the industry. 

. Physical Construction Lighting 

Leonard Satz Lester B. Isaac 

Century Circuit Loew's, Incorporated 

Auditorium Design Safety and Maintenance 

John Eberson Henry Anderson 

Architect Paramount Pictures 


Floor Coverings Television Projection 

A. Griffin Ashcroft Paul J. Larsen 

A lexander Smith and Sons Carpet Company Consultant 

Television Acoustics 

Donald E. Hyndman Harvey B. Fletcher 

Eastman Kodak Company Bell Telephone Laboratories 

Ventilating and Air Conditioning; 
Promotional Display 

Seymour Seider 
Empee Construction Corporation 

One of the high lights of the Conference was a demonstration of 
large-screen television by the Radio Corporation of America, attended 
by over five hundred people. 

The success of the Conference was indicated by record attendance, 
including a generous cross section of theater men and architects from 
all parts of the United States and twenty representatives from foreign 
countries. The Society, recognizing the tremendous interest in 
Theater Engineering, now offers the entire motion picture industry 
an opportunity to read or review all of the technical papers and the 
ensuing discussions as they took place at the conference. 

James Frank, Jr. 

Theater Conference Chairman 

Effect of Television 

on Motion Picture Attendance* 


Summary The purpose of this study is to obtain an indication of the 
effect of television upon motion picture attendance habits. Such a study 
could suggest the extent to which television will affect box-office receipts 
when set ownership has become more widespread than it is at present. 
Telephone numbers of 550 owners of home television sets were selected 
at random from the four major boroughs of New York City. Interviews 
were completed with 415 owners of sets presently in working order. 


FOR A STUDY of this type, telephone interviews seemed most ap- 
propriate, since they provided the double advantage of speed 
and economy. A recent report by The Pulse, Inc., shows that 87 
per cent of the set owners in New York City have telephones, indi- 
cating that an adequate sample of set owners could be reached by 

With a list of 10,000 set owners as a base, the telephone numbers 
of 550 were selected completely at random. The study was limited 
to the four major boroughs of New York City: Manhattan, Bronx, 
Brooklyn, and Queens. 

Interviews were completed with 415 set owners. Interviews could 
not be completed with the remainder of our list for various reasons : 

No longer had a set or temporarily out of order 13 per cent 

Refused (too busy, ill, unable to speak English) 4 per cent 

No answer 7 per cent 

In a study of motion picture going conducted at home, it is im- 
portant to keep to a minimum the number of people who are lost 
because they are not at home when called. In general, these are 
more active people-^and likely to be frequent patrons of motion 
pictures. If this study were limited to a single attempt to reach each 
family, some bias would be introduced into the sample. For this 
reason, four and five "call backs" were made in every case where 

* Presented May 17, 1948, at the SMPE Convention in Santa Monica. 


The resulting 7 per cent "no answer" is 
considered satisfactory for the purpose of the present study. 

The questionnaire used in this study is shown below. 


Good (Morning) (Afternoon) This is the General Research Bureau. We are 
conducting a survey among television-set owners. 

1. Do you have a television set at home? Yes ( ) No ( ) 

2. Is it in working order? Yes ( ) No ( ) 


3. How long have you had it? 

4. Do you think your television set has had any effect on either increasing or de 
creasing the number of evenings you spend at home? 

Increased ( ) Decreased ( ) Had No Effect ( ) 

4 (a) About how many evenings a week more do you spend at home? 


4(b) About how many evenings a week less do you spend at home? 

5. Let's take the movies as an example. Since you got your television set, do 
you think you go to the movies more often or less often, or is there no difference? 

More often ( ) Less often ( ) No difference ( ) 
5 (a) How often do you go to the movies now? 

times a week or times a month DK ( ) 

5(b) About how often did you go to the movies before you got your television 

set? -. -, 
times a week or times a month DK ( ) 


PHONE NUMBER: SEX: Male ( ) Female ( ) 



Some thought was given to the possibility that the structure of this 
questionnaire may have biased the results of this survey. It might be 
argued that the introduction and the wording of the questions would 
lead television-set owners to feel obliged to report a change. 

With this in mind, check interviews were made with a sample of 
people comparable to those included in the study itself. With these 

14 AUSTRIAN January 

people, no mention was made of television. Rather, the same ques- 
tionnaire was used with the substitution of "new radio" for "tele- 
vision set." Respondents in the check interview w r ere asked to re- 
port if there had been any changes in their motion picture attendance 
since they bought their new radios. 

These check interviews did not elicit any reports of gross changes 
in picture attendance. In fact, it was difficult for people to grasp 
why there should be a change. 

It seems reasonable to conclude, therefore, that there is no bias 
inherent in the structure of the questionnaire or the wording of the 
questions. If there is any tendency which leads to exaggerated 
answers, it lies in the fact of television-set ownership itself. But there 
is no reason to believe that the answers found in the study give an 
incorrect picture of the trend. 


Everything connected with television is changing rapidly from day 
to day. This study reflects present conditions ; it is not presented as 
a prediction of future developments. 

As television programs improve, the medium is likely to provide 
increasingly stiff competition for the motion picture producer. In 
interpreting the results of this study, the dynamic state of television 
should be kept in mind. 

Although most of the sets owned by. the people interviewed in this 
study were bought in the last year or two, many of them date back 
to before the war. It is perhaps too much to expect that a housewife 
who has had a television set for five or six years could give an abso- 
lutely accurate report of her motion picture attendance habits before 
she bought the set. What she tells us is not what actually happened 
(which is subject to errors of memory), but only what she recalls has 

Since the end of the war there has been a general decline in at- 
tendance at motion pictures. It is reasonable to assume that this 
has affected both set owners and nonowners. Interpretation of any 
reports of changes in attendance made to us in this survey must be 
tempered by knowledge of this general trend. 

This survey is limited to home-set owners. Thus, any effects of tele- 
vision reported here may be an underestimation because no account is 
taken of the effects on nonowners who view television at the homes of 
friends, in bars, and other public places. 



Television has had a definite social impact on the families inter- 
viewed. Three quarters of them report that they spend more eve- 
nings at home now that they have a set. 

This effect has extended to motion pictures. Half of the set owners 
interviewed report that they go to the movies less often after buying 
a set. 

Most of the people who are going to the movies less were formerly 
very heavy goers. The movies are losing some of their best 


Half of the people interviewed (51 per cent) report that they attend 
motion pictures less often since purchasing a television set. 

The remainder state that television has had no effect on their 
movie-going habits (except for three people who state that they go 
more often now) . 

Table I shows a comparison of the findings obtained in this study 
with the results of .a survey conducted in the fall of 1947 in the Los 
Angeles area by Television Research, Inc., of that city. Notice the 
similarity in results. 


F. C. & B. 

TV Research 

New York 

Los Angeles 

Per Cent 

Per Cent 

Attend less than before 



Attend about the same 



Attend more often 




Fifty-one per cent of the people interviewed in the present study 
reported a change in movie attendance. In what direction was this 
change made? 

Table II shows a consistent shift to less-frequent movie at- 
tendance. The majority of persons who reported a change appear to 
have been very heavy movie-goers before they got their set. Their 
attendance has dropped from an average of "once every few days" to 
an average of somewhat less than once a week. 

16 AUSTRIAN January 



As Reported by the 51 Per Cent Whose Habits Have Changed with Television 

(211 People) 

Before TV 

After TV 

Per Cent 

Per Cent 

Every few days 
Once a week 



Every two or three weeks 





Notice that scarcely any (only 3 per cent) of these set owners were 
infrequent movie-goers before buying a set ; afterward, fully 29 per 
cent fell into this category. The behavior of those who report a 
change in their movie-going habits deserves more detailed study. 

Table II shows that the majority of set owners had been very 
heavy movie-goers before buying their television sets. 

Fully 90 per cent (that is, 57 plus 33 per cent) used to go once a 
week or oftener, according to their own statements. Now consider 
the habits of these heavy movie-goers after buying a t el vision set. 


Reported by Those Who Used to Attend Frequently 


Every Few Days 

Once a Week 

after Television 

Per Cent 

Per Cent 

Every few days 


Once a week 


Every two or three weeks 



Once a month 








In Table III, the column showing the post-television habits of those 
who used to go "every few days" shows that most of these people have 
now dropped to only once a week in attendance. An additional 27 
per cent (all of whom used to go every few days) now go even less 
than once a week. 


As for those who used to go once a week, most of them now go only 
once every two or three weeks. The rest go even less frequently. 

It is reasonable to assume that television would influence other 
forms of social behavior as well as motion picture attendance. In 
order to get a better understanding of the effect of television upon 
home set owners, all members of our sample were asked whether or 
not television had affected the number of evenings they spend at home. 
Three fourths of the people we talked to reported that they spend 
more evenings at home since buying their sets. The remainder 
said that television has had no effect upon their habits in this respect. 

To what extent has the increase in "staying at home" affected 

Of the 75 per cent who said evenings at home have increased, 
63 per cent said they attend the movies less often. 

It seems reasonable to assume that if these persons are staying 
home more and attending the movies less, television has had con- 
siderable impact on their social life. Motion pictures seem to have 
been hit hard by this increase in "stay-at-home" habits. 

Some people, who tend to minimize the impact of television, have 
advanced the theory that although movie-going may fall off when a 
set is new, attendance will pick up again as the novelty wears off. 

These assumptions were not borne out by the data accumulated 
in this study. The age of the set did not appear to have any re- 
lationship to reported changes in movie-going. 

We must remember that television is still comparatively new, and 
it is still too early to judge the reaction of set owners as they become 
accustomed to this medium. 


As this study neared completion, a check study was conducted by 
Dr. Thomas Coffin of Hofstra College, Hempstead, New York, work- 
ing in co-operation with Foote, Cone, and Belding. 

The check study was conducted by personal interview. Inter- 
views were completed with 270 families in and around Hempstead. 
Of this number, 135 owned home television sets; the other 135 fami- 
lies were similar in all respects except that they did not own television 

The interviewers asked set owners the same questions as did Foote, 
Cone, and Belding in the telephone survey reported here. 


All families, as well, were asked for specific reports of entertain- 
ment activities such as attendance at movies, sports activities, radio 
listening, and so on, during the preceding week. 

Results of the interviews provide support for Foote, Cone, and Bel- 
ding's findings. In the personal interviews, 58 per cent report a de- 
crease in movie attendance. (F. C. & B. reported 51 per cent). 

The data on actual attendance at movies and other entertainments 
provide some validation of these findings. Most significant is the 
finding that set owners actually bought 20 per cent fewer movie tickets 
during that week when compared with the matched sample of non- 

Navy Photography in the Antarctic* 



Summary The primary purpose of Operation "Highjump," 1946-1947, 
was to train personnel, test equipment, and improve operational techniques 
in subzero temperatures. Every phase of the operation and the performance 
of equipment undergoing tests were photographed in color motion pictures 
with a view toward producing technical and training films for educational 

The many difficulties inherent in photographic operations in subzero tem- 
peratures and polar regions require special techniques. These and the 
malfunctions of cameras, and causes and suggestions for improvements are 
treated. The Navy is developing cameras more suitable for use in frigid 

parted for the Antarctic on Operation HIGHJUMP in December, 
1946, with it went sixty-eight Navy, Army, Marine Corps, and Coast 
Guard photographers. The primary objective of the Operation was 
to train personnel and test equipment in subzero temperatures. The 
secondary objective was to ch#rt and photograph little known, or 
unknown areas. The duty of the photographers was to photograph 
every phase of the operation in motion pictures and still photographs 
both in color and black and white. The photographs and motion 
pictures were to be used for documentary and technical records, 
training films, and to teach personnel cold-weather techniques. The 
time permitted Operation HIGHJUMP in the Antarctic was very short. 
Ships could not remain in those icy waters longer than two months 
because of the danger of becoming icebound and being crushed by 
giant ice floes when winter began. In that short period over 241,000 
feet of motion picture film were exposed, 109,327 aerial and still 
photographs taken and processed, and the necessary prints made. 
There was no special attempt made to conduct tests on photo- 
graphic equipment in the strict sense of the word. Many different 
types and models of cameras were used to record tests, experiments, 
and the over-all Operation as events occurred. There was no time to 
stage any action; it had to be photographed as it happened. Thus 
from a military standpoint the photographers received valuable 
training and learned many new techniques as they went along. No 
better operational cold-weather test could have been conceived for 
* Presented May 21, 1948, at ,the SMPE Convention in Hollywood. 





the cameras. Hourly records were kept of temperatures. Records 
were kept of camera malfunctions and there were many. 

It is known that it gets colder in some parts of the United States 
than the temperature encountered by Task Force SIXTY-EIGHT Opera- 
tion HIGHJUMP. Also some will recall that they have successfully 
made motion pictures in their own home community in temperatures 
colder than 50 degrees Fahrenheit without experiencing undue diffi- 

Fig. 1 Navy photographer and camera suspended on a cargo platform from 
a crane of the USS Burton Island to obtain scenes of the Navy Icebreaker 
as she progresses through the ice pack in the Ross Sea. 

culty with their cameras. The answer to this is that it takes many 
hours to chill a camera thoroughly. On an operation such as the 
South Pole area expedition, cameras remained exposed to the cold 
and, as a general rule, the ones intended for outside work are never 
brought in a warm building. There is a very good reason for this. 
After a camera has been thoroughly chilled in very cold temperatures 
and is brought into a warm place, every portion of it becomes 
thoroughly wet from condensation, even between the lens elements. 
Prolonged practice of this nature will eventually result in corrosion 




and rust if the camera is not completely dismantled and every part 
dried each time this occurs. It is very doubtful if a person in his 
right mind would stay out in subzero temperatures in his home en- 
vironment and take pictures for 
very long at a time. The 
chances are he would prepare 
his cameras, then dash out and 
spend a few minutes taking pic- 
tures, and return to warmer 
quarters. This could not pos- 
sibly be considered as a good 
cold-weather test for cameras. 
It must be remembered that a 
soldier, sailor, or marine who is 
fighting in subzero temperatures 
cannot be expected to find a 
warm building either to defrost 
his camera or himself. 


All cameras were completely 
delubricated and relubricated 
with cold-weather lubricants and 
tested in cold chambers. Close 
tolerances of working parts were 
made larger where possible. Yet 
failures occurred on all motion 
picture cameras from plus 15 de- 
grees Fahrenheit to minus 27 
degrees Fahrenheit. 

It is quite apparent now that 
the reason our cameras passed 
cold tests in the laboratory and 
would not function properly in 
the field in much warmer tem- 
peratures, is that they were not 
completely chilled during laboratory tests. Cameras should remain 
in cold chambers, with film loaded until thoroughly chilled. As a gen- 
eral practice, most cold tests are only for a few hours. The total time 
to chill a camera thoroughly depends upon the size and mass of metal 

Fig. 2 The author skiing to camera 
location with a heavy motion picture 
camera and storage battery. This 
illustrates the necessity for a more 
portable camera for cold-weather opera- 
tions. It also indicates that all hand- 
held cameras had frozen up and failed. 




in its construction. It is important that the film be chilled too. Film 
becomes very brittle in subzero temperatures, and sometimes this 
causes malfunction. It is conceivable that in some instances it would 
require twenty-four or more hours to chill a camera thoroughly. 

Electric-powered cameras as a class are more reliable than spring- 
driven cameras for subzero temperatures. Standard 35-mm motion 

Fig. 3 Navy photographers about to leave on a photographic 
assignment are given last-minute instructions. 

picture cameras were used with two types of 24-volt electric motors, 
ball-bearing and sleeve-bearing type. The sleeve-bearing type slowed 
to half speed at +15 degrees Fahrenheit. The ball-bearing type 
functioned satisfactorily as low as 27 degrees Fahrenheit. The 
spring-driven cameras all failed on the plus side of the Fahrenheit 
scale after a few hours' exposure to low temperature. 

Successful motion pictures were obtained from the air at tempera- 
tures of 40 degrees Fahrenheit with small 16-mm spring-driven cam- 
eras by keeping a stream of warm air sprayed on them. This would 
be impossible to do by a man operating a camera on the ground. 


The large studio-type cameras were cranked by hand when the 
electric motors failed. Having been in the Antarctic previously 
with Rear Admiral Byrd, the officer in charge of all photography 
during Operation HIGHJUMP required all photographers to become 
proficient in hand-cranking cameras. It was expected that elec- 
tric motors would fail. The average small, hand-held, spring-driven 
camera cannot be hand-cranked satisfactorily in extreme cold. 

The reasons for failures of instruments in subzero temperatures are 
not new. It is well known that the differential of thermal expansion 
and contraction of moving parts of different metals is the cause. In 

Fig. 4 Photographers hauling cameras on a sled at Little America. 

cameras the usual causes are moving steel parts being frozen by the 
contraction of aluminum or magnesium alloys. The contraction of 
these light alloys is much greater than steel. Consequently, if mov- 
ing parts housed in these alloys are not properly bushed with steel 
bushings and ball bearings and have proper tolerances, failures will 
occur at low temperatures. Tolerances can be increased on almost 
any camera so that it will operate at subzero temperatures, but when 
it is returned to moderate temperatures these tolerances usually be- 
come so great that the camera will be useless until it is reworked and 
the tolerances reduced. Navy specifications require that all cameras 
operate in any temperature likely to be encountered. 

In the past the Navy has, except for special-purpose cameras, 
generally procured cameras that were basically engineered for the 
commercial market. A few changes were generally specified but 

24 SHIRLEY January 

these were usually minor. Cameras produced for the commercial 
market are not built to operate in all the temperatures required by 
the Navy. There are very good reasons for this. Such a camera, 
one which will operate in a temperature range of from 67 to +141 
degrees Fahrenheit, probably can be produced, but the expense would 
be prohibitive for a commercial camera. All Navy cameras should 
not be expected to functon under these extreme conditions. In the 

Fig. 5 A weasel equipped for photography at 
Little America. Large, heavy cameras were made 
portable in this manner. 

event of another war, it is reasonable to believe that extreme tem- 
peratures will be encountered by some portion of the Navy almost 
every day, and cameras, like other instruments, must not fail. The 
contention is that a great many expensive cameras engineered for the 
above-mentioned temperature range may neve.* be used in colder 
temperatures than freezing or higher temperatures than experienced 
in the tropics. A more reasonable solution, and a less-expensive one, 
would be to produce special subzero-weather cameras, required to 
operate from 70 degrees Fahrenheit up to freezing, and temperate- 
weather cameras for a range of about 10 up to 141 degrees Fahrenheit. 



Film is a great source of trouble in subzero temperatures. It be- 
comes very brittle and will break easily. Sharp bends must there- 
fore be avoided. Research is required to find a more suitable plastic 
for film bases for subzero weather than is currently used. It may be 
that this research will prove that the present type of emulsion is 
causing most of the trouble. The differential of thermal expansion 
and contraction between the plastic base and the emulsion is no 
doubt the cause of some of the reactions of film in subzero tempera- 
tures. That is probably why film is inclined to curl very tightly to- 
ward the emulsion in cold temperatures. It is doubtful if industry 

Fig. 6 Filming scenes of O. F. Bo we as he progresses through a 
dangerously crevassed area on the shelf ice west of Little America. 
The safety line around the author is attached to a weasel. 

has ever been called upon to produce film that will be pliable in ex- 
tremely low temperatures. There is very little commercial need for it. 
Static is also a great source of trouble in extremely low tempera- 
tures. Research should be conducted to try to reduce or eliminate this. 


The friction-head variety is undesirable in subzero temperatures. 
Smooth operation even in mild climates requires that the head be 
packed with grease. The grease has to be replaced with a lighter 
lubricant of oil and graphite for cold-weather operations. With the 
heavy grease removed, the head wobbles and jerks when panned. 
Gyro-type or cranked-head tripods are probably satisfactory but the 
Task Force had none of these. 





Much was learned from experience during Operation HIGHJUMP 
that may be helpful to anyone taking pictures in cold weather. 
Some of these techniques are listed below. 

1. Never breathe on a lens; it will cause condensation which 
freezes instantly. The resulting ice is very difficult and sometimes 
impossible to remove unless the lens is taken in to a warm place. 

Fig. 7 Photographer dressed in cold-weather gear on board the 
USS Mount Olympus, Flagship of Rear Admiral Richard H. Cruzen, 
Task Force Commander. 

2. Never attempt to clean a lens with an ungloved hand. Body 
heat will be transmitted to the lens and cause it to ice over. Avoid 
breathing on the eyepieces and viewfinders for the same reason. 

3. Keep the eye far enough from the eyepiece so that body heat 
will not be transmitted to it. There is danger of the eyelids' sticking 
to an "eyepiece if they come in contact with unpainted metal. 

4. Avoid touching unpainted metal surfaces with the bare skin. 
Painful injury will result, especially from touching unpainted steel. 
The skin will stick to it as if glued. A portion of skin is always lost 
if one is careless about this. 




5. Do not take a thoroughly chilled camera from the cold to a 
warm place with the intention of using it immediately. The camera 
cannot be used until its temperature equals the surrounding warmer 
temperature. Several hours are required as a general rule for the 
camera and lenses to complete collecting moisture from condensation 
and thoroughly dry. If a camera is removed to the cold before be- 
ing completely dry, icing will result. This can be serious if it occurs 
on the iris diaphragm or in some interior moving part of he camera. 
Keep cameras to be used in extreme low temperatures stored out- 
side and those for interior work indoors in a warm temperature. 

Fig. 8 Loading a motion picture camera barehanded at Little 
America, a painful task in subzero weather. Note the crank in 
place which indicates that the camera had to be hand-cranked 
because the electric motor had failed. 

6. Never take a warm camera out in a blizzard with the expecta- 
tion of getting good pictures. Drifting snow striking the lens will 
melt, and in a very short time the lens will be covered with ice. 


A motion picture camera for subzero photography for combat use 
should embody the following features : 

1. It should be light in weight and portable. 

2. It should be as void of unpainted surfaces as possible, especially 
those surfaces which have to be touched or adjusted with bare hands. 




3. It should be semiautomatic magazine load, regardless of 
whether it is 35-mm or 16-mm. One of the most painful things 
imaginable is threading a 35-mm studio-type camera barehanded in 
temperatues of 50 degrees Fahrenheit or even more moderate tem- 
peratures. It is possible to use skintight gloves which will prevent 
considerable loss of the skin, but it is impossible to thread such a 
camera with the hands adequately gloved for warmth in subzero 
temperatures. One of the photographers on Operation HIGH JUMP 
became very proficient at loading and threading a, large studio-type 
camera with fully mittened hands by using a pencil to adjust the film. 

4. All adjustments should be possible with the hands encased in 
three heavy woolen mittens plus an outer leather mitten shell. 

Fig. 9 Photographic operations headquarters during Operation 
HIGHJUMP at Little America, Antarctica. Camp and living quarters 
may be seen in background. 

5. Finders and eyepieces, if employing lenses (except for focusing, 
lenses are not recommended), should be well insulated with rubber and 
well ventilated with holes between the rubber and lens so that body 
heat will escape before it reaches the glass surface of the lens . 

6. Electric power is preferred. The power source should be com- 
pact and capable of being strapped to the body in such a manner that 
it will not handicap movement. 

7. _ Hand-held-type cameras are preferable and should be equipped 
with a shotgun-type stock and trigger. 

8. A choice of three to four taking lenses and finder lens, mounted 
in a turret that can be rotated easily and positively locked. 

9. It should function smoothly at 67 degrees Fahrenheit. 



An entirely unexpected phenomenon encountered in the Antarctic 
is the abundance of reflected light. This is surprising to one who 
follows the practice of increasing exposure as the distance north or 
south of the equator is increased. In preparation for an expedition 
to the Antarctic, the average photographer would assume that a very 
fast emulsion would be required. As a general rule, motion picture 
film faster than Weston 50 is not usable without neutral density or 
other exposure-reducing filters. 

Another surprising phenomenon in the Antarctic is the fact that 
there is more light on overcast days than on bright, clear, cloudless 
days. This is explained by the fact that practically all of the light 
striking the surface is reflected. Upon striking the bottom of the 
overcast it is reflected back to the ice surface. A continuation of this 
reflection back and forth causes a build-up of effective light. Thus, 
the Antarctic is probably the only place in the world that requires less 
exposure on overcast days, except that it is reasonable to believe 
a similar situation exists in the Arctic. 

Because of the abundance of light in Antarctica it is necessary to 
mask off a portion of the photocell aperture on the average exposure 
meter to obtain a reasonably accurate reading. On two of the most 
popular type of meters the indicating needle registers beyond the high- 
est calibrations. By masking off 50 or 75 per cent of the cell aperture 
and multiplying the reading obtained by the percentage mask off, 
fairly accurate exposure calculations may be obtained. 

Much can be written about ice photography. Probably the most 
authoritative work on ice photography is that by Herbert G. Ponting, 
who accompanied Captain Robert Falcon Scott of the British Royal 
Navy to the Antarctic as a photographer in theearly part of the 

1. For best detail on an ice surface, the picture should be taken 
against the light source. 

2. Detail is not possible on a flat ice surface when the view is 
180 degrees from the light source. 

3. The angle relative to the light source can be determined by 
observations from several positions, and will vary in accordance with 
the results desired. 

4. A good "Rule of Thumb" for exposure is to expose for detail 
in shadows. 

All illustrations are Official United States Navy photographs. 

Motion Picture Photography at 
Ten Million Frames per Second* 



Summary A new procedure is used in which the image of a rectangular 
picture is broken up and reassembled into a long narrow strip. After process- 
ing, the negative must be reconstructed into a rectangular motion picture 
frame by projection printing through an optical system similar to that which 
formed it. 

To EXPOSE A conventional motion picture at a speed of several 
million frames per second would require a speed of film move- 
ment of the order of 200,000 feet per second for 16-mm film, a rate 
entirely beyond anything attainable at present. As an alternative the 
image may be swept across stationary film at speeds much higher than 
it is possible to move the film itself, but this procedure imposes a very 
serious limit upon the length of film and hence the number of frames 
which can be exposed. 

The required film or image speed could be much reduced if the con- 
ventional shape of the motion picture frame were altered to make the 
dimension along the direction of film movement very small. If at the 
same time the dimension across the film were increased a correspond- 
ing amount to preserve the same picture area, the same total number 
of just-resolvable .picture elements might be retained. In the camera 
to be described this change of image shape is accomplished by a 
stationary optical system through which the motion picture negative 
is exposed. After processing, the negative film is printed by projection 
through a similar optical system which reconstructs the image back to 
the shape of the original object and thus to the approximate propor- 
tions of the normal rectangular motion picture frame. In this manner 
motion pictures in excess of ten million frames per second have been 
produced with very moderate film velocities, the arrangement per- - 
mitting the photography of a very large number of motion picture 
frames in a single sequence. By the use of an automatic printer the 
final positive is presented as a standard 16-mm motion picture print. 

* Presented October 21, 1947, at the SMPE Convention in New York. 



The optical system changes the shape, of the picture by cutting the 
original image of the object into a series of narrow strips, redisposing 
these strips end to end, and reimaging them as a single long strip 
extending across the motion picture negative film. We have termed 
this process image dissection, and the optical system which accom- 
plishes it an image dissector. The basic idea of cutting an image into 
strips is not new, and methods for accomplishing this have been 
described by Walton. 1 However, to meet the requirements of ultra- 
speed photography it has been necessary for us to devise a new type of 
optical system. 

The principle of the image dissector is shown in Fig. 1. Consider a 
number of small identical objective lenses L, as shown in Fig. 1, with 
their optic axes perpendicular to 
the plane of the paper. For con- 
venience these lenses are cut with 
flat edges and blocked together 
as shown. If the line joining 
their centers is slightly inclined 
to the horizontal, and a distant 
event is viewed by these lenses, 
each will form a separate but 
identical image of the distant 

Fig. 1 A multilens unit L and its 
images on a slit S, illustrating the 
method of image dissection. 

event, and these images will ap- 
pear as a flight of steps as shown 
in the upper part of Fig. 1. 
If these images are received on a metal plate containing a narrow slit 
S, which is not inclined but truly horizontal as shown in the figure, it 
is evident that this slit will pick just the top of the picture from the 
first image, a strip of the picture next below from the second image, 
next below that for the third image, and so on. Since all the images 
are alike, it will be evident that the various parts of the slit from left to 
right carry the equivalent of one complete picture of the event. This 
we have referred to as image dissection, and in the example shown in 
Fig. 1, the rectangular image is dissected into five narrow strips which 
are assembled end to end to form one long narrow strip which passes 
through the slit. If this narrow strip is imaged on photographic nega- 
tive film which is moving in a direction from top to bottom referred to 
in Fig. 1 , it will be seen that vertical streaks will be formed on the film, 
their position corresponding to the light and dark portions of the origi- 
nal picture. If the slit is made sufficiently narrow, and the number of 




lenses sufficiently great so that the whole picture is represented en- 
tirely in this narrow slit, then the blurring of the picture which results 
from the motion of the film cannot be greater than the width of the 
slit, and this in turn can be made as small as the finest detail which the 
photographic negative is capable of resolving. While such a streak 
negative could be analyzed and the necessary information secured 
from it for many scientific purposes, it is far more convenient to print 
the negative back on to positive film by projection through the very 
optical system which formed it. In this manner the picture is rectified 
from streak images back to an ordinary frame of motion picture film. 
Each position of the slit across the negative film gives a new frame of 
the final motion picture. 

Fig. 2 Optical arrangement of a 5-element image dissector as used for near 


A schematic diagram of the complete optical system is shown in 
Fig. 2, including the provision for photographing an object which is 
not at infinity. Light from the object at is collimated by the first 
objective lens LI, and then received by the multiple objective lens 
system L/%. These lenses form the multiple images in the plane of the 
slit S. The image of the slit S is formed on the film at F by the main 
lens Ls, which must be a very well-corrected photographic objective. 
In order that the pupils of the multiple optical system shall fall on the 
final objective L^ two sets of specially designed field lenses, L 3 and 
L 4 , are mounted very close to the plane of the slit. By this arrange- 
ment the effective photographic speed of the combined optical system 
is equal to the photographic speed of the lens L 5 , subject only to sur- 
face reflection losses of the lenses LI to L 4 , inclusive. These reflection 
losses can be made quite small by suitable nonreflection treatment. 
In the present camera the final lens L$ is a photographic objective 
operating at //2.O.* 




In our Model I camera there are fifteen small objectives L 2 instead 
of the five shown in Fig. 2. When used with a single slit at S this 
produces a final picture having a total of fifteen elements only. To 
increase the number of elements multiple slits may be used. Refer- 
ring again to Fig. 1, if the inclination of the line joining the cen- 
ters of the lenses L be reduced to one half, and if the slit width be re- 
duced to one half, then it is evident that only the upper half of the 
original picture will be covered by the five slit elements. Suppose now 
that a second slit be placed parallel to S but spaced below it by just 
five times the width of either slit. Under these conditions with a single 
set of five lenses L, the number of elements of the final picture will be 

Fig. 3 Fifteen-element image dis- 
sector. The slit jaws may be seen as a 
faint dark streak through the multi- 
condenser unit. Above this is a device 
for placing fiducial marks on the nega- 
tive. The multilens unit is largely con- 
cealed, but its focusing adjustment 
shows prominently (large knurled head 
in foreground). 

Fig. 4 Fifteen-element image dis- 
sector viewed from object side. The 
end of the multilens unit may be seen 
near the middle of the picture, and 
below it the thumbscrew for adjusting 

doubled, the first slit taking care of the upper half of the final picture 
and the second slit, the lower half. Photographs of the complete 
image dissector with single slit are shown in Figs. 3 and 4, without the 
final photographic ob j ective L 5 . A photograph of the double-slit system 
is shown in Fig. 5. Although the two slits are very close together, their 
images on the final negative film must be widely spaced to avoid over- 
lap of the streak image. This is accomplished by the reflecting prisms 

* This lens is a Kodak Ektar of 45 mm focal length, designed for object at 
infinity. It is here required to work at 5 to 1 conjugates. F. E. Altman, of the 
Eastman Kodak Company, very kindly arranged for a supplementary lens sys- 
tem which is used in front of the Ektar to permit the latter to operate at its de- 
signed conjugates. 




Fig. 5 Image dissector with double 
slit in place. 

shown in the top and bottom 
right of Fig. 5, which are provided 
with small micrometer screws 
to adjust their position. With 
the arrangement shown in Fig. 5, 
a 30-element picture is formed. * 
In Fig. 6 is shown the manner 
of handling the negative film 
which was described some years 
ago. 2 A shallow drum, open at 
the top, is mounted on the 

Fig. 6 Cross section of high-speed rotor illustrating 
method of imaging slit on the moving film. 

shaft of a vertical high-speed motor. The inner circumference 
'of the drum is machined to accept one turn of 16-mm film cut 
with a gauge 24 inches long so that it just fits within the drum 
with a negligible gap. Light from the image dissector reaches the 
mirror MI, is reflected down through the main photographic ob- 
jective L, and reflected once more by the lower mirror M 2 , to form the 

* The small objective L 2 , although simple cemented achromats, must be of good 
quality. L. V. Foster, of the Bausch and Lomb Optical Company, very kindly 
arranged for the procurement of the optical elements of the Bausch and Lomb 40- 
mm microscope objective which proved very satisfactory. 




image on the rotating film. At top rate of speed this film is driven 
past the image at the rate of 400 feet per second. By proper attention 
to mechanical and optical detail it is possible to resolve approximately 
100 lines per millimeter on this film. At a resolution of 80 lines per 

Fig. 7 The complete camera with image dissector in place. 

millimeter with the film traveling 400 feet per second, the individual 
exposures are of one-ten-millionth-second duration, and in effect ten 
million separate motion picture frames per second are photographed. 
The fully assembled camera is shown in Fig. 7. The diameter of the 
rotor case is about 10 inches and the over-all height about 18 inches. 
The present complete camera weighs about 60 pounds, much of this 


weight being in the cast-iron base. The driving motor, rated at 12,000 
revolutions per minute, is shown immediately below the rotor case. 
Below the motor, directly connected to its shaft is a miniature induc- 
tor type of alternating-current generator. The frequency of the alter- 
nating-current output of this generator is a very accurate measure- 
ment of rotor speed, a matter of some importance in certain scientific 
studies with the camera. The main objective lens Ls is in the vertical 

Fig. 8 The printer arranged for automatic rectification 
of the streak negative (placed in the precision carriage at 
left) into a finished 16-mm projection print. 

column above the rotor, and the image dissector appears on the upper 
arm, the housing being removed to show it more clearly. 

When using an image dissector with a single slit, the 24-inch length 
of film will carry more than 60,000 slit images. Since each complete 
slit image constitutes a motion picture frame, it is evident that, if 
necessary, a single scene of more than 60,000 frames' duration can be 
photographed. In using a double-slit system giving a 30-element pic- 
ture, the number of frames which may be photographed without over- 
lap is much reduced. In the present model this is limited to continu- 
ous runs of 1600 frames, which is still sufficient for photographing a 
great variety of events. 




Since at a speed of ten million frames per second the exposure 
time is limited to one-ten-millionth second, the problem of get- 
ting sufficient light for proper photographic exposure can be very 
difficult. If one calculates the illumination required upon ordinary 
opaque objects of average reflectance, the value turns out to be be- 
tween twenty and one hundred million foot-candles. To obtain such 
an illumination upon even one square foot would require several thou- 
sand kilowatts of the most efficient light sources such as carbon or 
mercury arcs, and such power requirements are obviously out of the 

Fig. 9 Reassembled motion picture frames showing 0.125-microfarad 
condenser charged to 30 kilovolts discharged through 0.003-inch vertical 
iron wire. Magnification on print approximately six times, approximate 
intervals: A, second; B, 0.05 microsecond; C, 0.1 microsecond; D, 0.4 
microsecond; E, 1 microsecond; F, 3 microseconds. Note the initial expan- 
sion rate of over 10 kilometers per second. This expansion rate is actually 
too fast for proper resolution. 

question except in specially equipped laboratories. Fortunately mod- 
ern electrical-discharge flash lamps provide an ideal solution because 
the duration of the high-intensity illumination required is quite short. 
Suppose a scene sequence of 2000 frames is required at ten million 
frames per second. This means that the illumination must continue 
for only one-five-thousandth second. Flash lamps of the types de- 
scribed by Edgerton and his associates can easily produce a total flux 
of 10 9 lumens for times greater than 10~ 4 second. For example, a 
General Electric type FT 524 lamp operated from a capacitor of 100- 




microfarad capacitance charged to 4 kilovolts will produce the order of 
10 9 lumens for about 2 X 10 ~ 4 second, an operating discharge which is 
entirely practical in this lamp. A capacitor and power supply weigh 
only about 75 pounds, providing a practical and at the same time 
portable light source. A timing mechanism is, of course, necessary to 


Fig. 10 Reassembled motion picture frames of a 0.22-caliber rifle bullet 
passing through 2-mm glass plate. Photographed at 5,000,000 frames per 
second, prints reproduced at intervals of approximately 40 frames or 8 micro- 
seconds. The actual bullet is about 1 / 2 inch long. Note as the bullet strikes 
the plate a fracture wave travels vertically at approximately three times 
bullet velocity. After the bullet has passed a cloud of glass fragments remains 
"suspended" in mid -air, the forces acting on these, particles now being rela- 
tively small. . 

discharge the lamp at the appropriate moment required by the object 
to be photographed, but in most applications this is comparatively 
simple. Numerous timing methods have already been described by 

After the negative film is processed it is necessary to print it back 
through a system similar to that of the image dissector, in order to 






reconstitute the original picture. This is done with a simple auto- 
matic printer shown in Fig. 8. The negative film is mounted on a mi- 
crometer slide which is advanced automatically the required interval 
by means of a solenoid-operated ratchet. The reconstructed image is 
received on standard 16-mm positive film carried in the small motorr 
driven camera provided with a micros witch on the single frame shaft. 
This switch closes the solenoid circuit momentarily during the pull- 
down interval of the camera, advancing the negative by any desired 
amounts from one to twenty 
frames. Thus if the original pho- 
tograph has been made at a 
higher speed than is required to 
show the motion it is only neces- 
sary to print every second, fifth, 
or even every twentieth frame in 
making the final 16-mm film. 

In Fig. 9 is shown a sequence of 
the explosion of a metallic wire 
subjected to a heavy capacitor 
discharge. This is a very bright 
self-luminous event and presents 
no illumination problem. In Fig. 
10 is shown a photograph of a 
rifle bullet initiating a transverse 
fracture wave in a vertical glass 
plate. In Fig. 11 is shown an en- 
larged reproduction of the nega- 
tive from which Fig. 10 was made. 
It will be noted that the upper 
half of the photographs of Fig. 
10 have been supplied by the 
upper portion of Fig. 11, while the lower halves of each picture have 
been provided by the lower half of Fig. 11. A single frame of the 
final motion picture is represented by a very narrow strip extending 
from left to right all the way across the upper portion of Fig. 11, 
plus a similar narrow strip extending across the lower portion of the 
figure. With the film speeds used at present this strip on the origi- 
nal negative is only one-eightieth millimeter high with the camera run- 
ning at ten million frames per second. 

The most conspicuous feature of the present result, other than high 

Fig. 11 Negative streak image of 
rifle bullet breaking glass panel. The 
bullet is entering the field at A, the 
fracture wave is commencing at B, and 
the bullet leaves the field approxi- 
mately at the position C. 


speed, is the very poor image quality. Although the camera is useful 
in its present form for the analysis of certain types of fast events, the 
poor image quality is a very serious practical limitation. This is fully 
recognized, and this paper is in the nature of a progress report. 
Another form of the camera giving similar speed but much better 
image quality is now under construction, and it is hoped that a report 
upon this may be made at an early date. 



(1) U. S. Patents Nos. 2,021,162 (1935); 2,061,016 (1936); 2,088,732 (1937); 
2,088,626(1937); 2,089,155(1937); 2,112,002(1938); assigned to Scophony Ltd., 

(2) "High-speed running film camera for photographic photometry," Phys. 
Rev., vol. 50, p. 400; 1936. 

Visual Test Film 

THE 35-MM VISUAL TEST FILM first announced jointly by 
the Society of Motion Picture Engineers and the Motion 
Picture Research Council in January, 1947, is now available on 
safety base only. The price per roll is $22.50 postage paid in 
the United States, except in New York City where 45 cents 
must be added for city sales tax. When shipped outside the 
United States or possessions, the postage-paid price is $25.00. 

Recent changes in the method of printing have resulted in 
improved steadiness, making this film a reliable performance 
test for 35-mm projectors. Vertical unsteadiness is measured 
in per cent of picture height while horizontal unsteadiness is 
measured in per cent of picture width. Unsteadiness values as 
low as one fourth of one per cent may be determined readily. 

Screen masking may be adjusted and projector alignment 
checked with the use of the Focus and Alignment target. A 
Travel Ghost target provides a sensitive indication of shutter- 
timing errors and a series of vertical and horizontal lines indicate 
lack of sharp focus or curvature field. 

Complete instructions for use are supplied with this film. 

Comparison of Lead-Sulfide 
Photoconductive Cells with 
Photoemissive Tubes* 


Summary A comparison is given for lead-sulfide photoconductive cells 
versus photoemissive tubes with SI and S4 response with respect to signal 
(expressed as voltage sensitivity dV/df) at different color temperatures of the 
exciting light source. Spectral response, linearity, uniformity, and frequency 
response of lead-sulfide cells are also discussed. 

At present most of the phototubes used in sound reproduction are caesium- 
oxide gas-filled tubes. Of late, however, following a publication by R. J. 
Cashman 1 considerable interest has been shown by the motion picture 
industry in the lead-sulfide cell. In the past year some additional insight 
has been gained as to the problems arising with the application of these cells 
to sound reproduction and it seems that the motion picture industry should 
be apprised of some of these problems in order to design their future equip- 
ment. In this paper we shall endeavor to discuss the lead-sulfide cells as 
compared to photoemissive tubes and to point out some modifications of the 
present practice which have to be followed in order to get optimum perform- 
ance from lead-sulfide cells. 


s SHOWN PREVIOUSLY, the lead-sulfide cells have a much higher 
response to infrared between 1 and 3 microns than the presently 
used photoemissive-type tubes. The spectral-response curve enters 
into consideration in several factors of the sound-reproduction system. 

A. Spectral response of the tube itself. 

B. Spectral distribution of the exciting-lamp source. 

C. Spectral distribution of the transmission and refraction char- 
acteristics of the optical glasses used in the optical system. 

D; Opacity to radiation of the sound tracks in different regions 
of the spectrum. 

The spectral response of the lead-sulfide surface is a variable and 
to some extent controllable feature in the production of such cells. 
It has been shown 2 that the spectral response depends greatly on the 
relative amounts of oxygen and sulfur in the sensitive layer. In 

* Presented February 12, 1948, at the Midwest Section Meeting in Chicago. 







u- 0.66 


Fig. 1, curve A shows the response of a cell containing very little lead 
oxide. Curve C shows the spectral response of a cell with a high 
content of lead oxide, and curve B gives the spectral response of an 
intermediate-type layer. Another factor influencing the spectral 
response seems to be the thickness of the coating. It can be expected 
that the final shape of the spectral-response curve for cells used by 

the motion picture in- 
dustry can to some ex- 
tent be adapted to values 
required by the three 
other factors. 

In Fig. 2 are repre- 
sented spectral-distribu- 
tion curves of a black 
body at different true 
temperatures in degrees 
centigrade. These figures 
are relative figures cor- 
rected to the same watt 
output and it is shown 
that the maximum shifts 
toward longer wave- 
lengths at lower tem- 
peratures. For use with 
a lead-sulfide surface, a 
source with a maximum 
between 1 and 2 microns 
seems proper. For this 
reason it has been sug- 
gested to use indirectly 
heated exciter lamps. 
Work is being done at 
present on such lamps. 
The final spectral-distribution curve of the exciter lamp will then 
include the emissivity factor of the radiating material and the color 
temperature at which these lamps can be run without decomposition 
or deterioration of the radiator. 

There are few data available on the infrared transmission of 
crown and flint glasses. Some published data 3 show that for Jena 
glasses the coefficient of transmission of crown glass starts to decrease 


1.0 2.0 3.0 


Fig. 1 From L. Sosnowski, et al., reference 2. 
Reprinted by permission of Nature Magazine. 

Curve A Small amount of lead oxide. 
Curve B Intermediate. Curve C High con- 
tent of lead oxide. 


considerably between 1.7 and 2.5 microns depending on the type of 
glass. Flint glasses seem to be transparent to somewhat longer 
wavelengths. This transmission as well as the refraction characteris- 
tics leading to the possibility of designing chromatically corrected 
systems requires further study. 

The same consideration is true for the opacity of different film 
materials in the infrared. 

Summing up, it should be pointed out that for practical applica- 
tions in the future a certain standardization on the type of spectral 
response of the lead-sulfide cell would be of interest. 


A paper was published by A. Cramwinckel 4 in which different types 
of photoemissive tubes were compared, with relation to their sensitiv- 
ity, to photovoltaic selenium cells. It seemed to us of interest to 
extend the comparison to lead-sulfide tubes. Because of the differ- 
ence in mechanism of photoconductive and photoemissive types a 
common basis of comparison had to be found. As such the " voltage 
sensitivity" as shown in the book of Zworykin and Wilson 5 seems to 
give the most comprehensive results. "Voltage sensitivity" is the 
voltage developed across the series resistor per unit flux and its value 
is determined by the following equations: 


dV_ _ Rs dV _ ER-dr 

df ~ 1 + Rf(ds/de) df ~ (R +'r) 2 2df 

= Rs (vacuum tubes) , . Edr ... 

^ RsA (gas tubes R < 1 megohm) = 4- (optimum value when R = r) 

V = voltage developed across load resistor R 
R = load resistor 

/ = flux in lumens or watts 

s = luminous sensitivity of photoemissive tube in microamperes per lumen 
at some specified color temperature 

e = voltage developed across the photoemissive tube 
ds/de = change in s per unit change in voltage across photoemissive tube 
A = gas amplification of gas photoemissive tubes 
E = voltage supply in circuit 

r = resistance of photoconductive cell when exposed to a radiant flux / 

We have plotted the curves of voltage sensitivity using these equations 
(Fig. 3). Voltage sensitivity in volts per microwatts is plotted as a 
function of the color temperature of a calibrated tungsten lamp. 




Corrections have been made so that equal radiant energies are obtained 
for each color temperature. These corrections are tabulated in various 
books. The corrections used here were taken from Moon. 6 The cor- 
rection factor in going from 2870 to 1700 degrees Kelvin is about 12.3. 
For the 200-watt tungsten lamp used in obtaining these data, the 
difference between true temperature and color temperature is very 
small and hence may be neglected when using a color-temperature 
scale. If a lamp is calibrated in volts at a fixed color temperature, 
then the voltage that must be applied to'obtain any other color tem- 
perature can be calculated by a*formula given by Moon. 7 This was 

the method used here to 
obtain the range of color 
temperature given in the 

Curves 1 and 2 give 
the voltage sensitivities 
of the SI and S4 photo- 
emissive vacuum tubes 
with a load resistor of 
1 megohm. Curves 3 
and 4 give the voltage 
sensitivities of the SI 
and S4 gas tubes with 
a load resistor of 1 meg- 
ohm. The voltage sen- 
sitivity of the gas tubes 
is limited by the factor 
R (ds/de) which becomes 

20COO 30000 40000 


Fig. 2 Spectral distribution of the radiation 
from a black body. The wattage output in this 
case was kept constant. 

Curve A 2970 degrees Kelvin. Curve B 
2150 degrees Kelvin. Curve C 1500 degrees 

'quite appreciable as the 

load resistor is increased. However, if R is chosen as 1 megohm or 
less the voltage sensitivity of the gas tube becomes as many times 
greater as the gas-amplification factor A. For these data the gas 
amplification for the blue tube is taken as four and that of the 
red tube as eight. It must be noted that curves 8 and 4 give the 
maximum voltage sensitivity that may be expected by a load re- 
sistor of 1 megohm ; in practice it will be lower than that indicated in 
the graph. Curves 5 and 6 give the theoretical voltage sensitivity of 
the Si and S4 vacuum tubes with a load resistor of 100 megohms. 
In practice a high load resistor of this order can be used only when 
weak light levels are being detected. 




The rated current sensitivities of the photoemissive tubes used here 
are as follows: 

S4 vacuum tube 40 microamperes per lumen 
Si vacuum tube 20 microamperes per lumen 
S4 gas tube 160 microamperes per lumen (gas amplification 








^ ' 

-O.IOcm 2 
















5' q 




S-4 100 Meq load) VACUUM 









CE 1.3cm 2 

















*Ho* 3 







AD Ml 



r^ _q 





y^ 1.^* 




1 1 











Meq. 1 
















9 - 




. J 


". /T^\3l 
' V*J 

i 1 





- -- i 









in' 5 


1 1 1 




1900 2100 23OO 2500 7O0 


Fig. 3 Voltage sensitivity of different photoelectric surfaces as a function of 
the color temperature 


SI gas tube 160 microamperes per lumen (gas amplification 

The voltage sensitivity of the photoemissive tube is independent of 
area for a constant flux. However, on the lead-sulfide cell, the voltage 
sensitivity varies inversely with the area, again for constant flux. For 
this reason, various curves are plotted in which the area of the lead- 
sulfide surface is increased in the range 0.1 to 6.0 square centimeters. 
Data were obtained by taking an average of a number of cells having 
an area of 0.1 square centimeter. The curve showing the lead-sulfide 
surface of 1.3 square centimeters was calculated in order to make 

Fig. 4 Lead-sulfide cells. 

comparisons with the photoemissive gas tubes which are being used 
now in sound reproduction. The curve plotted for the 6-square- 
centimeter lead-sulfide surface was plotted for comparison with the 
photoemissive surfaces used in taking these data as the areas of these 
cells were 6 square centimeters. 

For comparison of these surfaces as to. color temperature let us use 
curves 8, 3, and 4- It is evident from these curves that the blue tube 
is increasing in voltage sensitivity as the color temperature increases 
toward 3000 degrees Kelvin. The red tube is increasing in voltage 
sensitivity also but at a much slower rate; however, the lead-sulfide 
surface is fairly flat in voltage sensitivity in this region of color tem- 
perature having a light peak near 2500 degrees Kelvin. At low color 


temperature of the order of 1900 degrees Kelvin the voltage sensi- 
tivity of the lead-sulfide cell is still fairly flat while that of the blue 
tube is decreasing very rapidly being down a factor of about 17 from 
that of the lead-sulfide cell and that of the red tube being down by a 
factor of 2.5. At lower color temperatures the voltage sensitivity 
of the lead-sulfide cell falls off quite rapidly and this would have to be 
taken into account in designing an indirectly heated exciter lamp as a 
source for the lead-sulfide cell. 

From these considerations it can be readily seen that the best ad- 
vantage of the lead-sulfide cell will be had when the optics are such as 
to permit the use of small-area cells of the order of 0.1 square centi- 
meter ( l /s X 1 /s inch). If a cell of this area is used the advantage in 
voltage sensitivity over the photoemissive type cell will be about a 
factor of 10. These curves for lead sulfide were calculated on a basis 
of a matching load resistor for each value of light resistance. In prac- 
tice this is not feasible and some loss is to be expected in voltage sensi- 
tivity for this reason. 


It follows from the discussions of the last paragraph that an ade- 
quate lead-sulfide cell should have a small area which would permit 
small over-all size of the tubes and would require particular care in the 
exact positioning of the light spot to the contacting pins or other ex- 
terior components of the tube. On Fig. 4 are given photographs of 
lead-sulfide cells which satisfy these requirements. These cells offer 
an advantage of uniformity in mass production, close tolerances of the 
sensitive area with respect to conductive pins, and direct insertion 
into a socket in which they can be held rigidly in place without the 
need of a base. This development is in line with the trend of the 
radio-tube industry to miniature and subminiature sizes. 


Photoemissive vacuum tubes are linear with light levels up to the 
order of 5 lumens and the response is independent of voltage if the 
voltage is maintained above that needed for saturation. Certain cells 
may be limited in linearity due to the existence of electrical leakage, 
fatigue, or an accumulation of space charge. Photoemissive gas tubes 
are linear with a light flux up to 1 lumen on the surface but not linear 
with voltage above 25 volts. Lead-sulfide cells of average sensitivity 
are linear with light levels up to 0.01 lumen on the surface. The 
linearity breaks off at lower light levels than this for very sensitive 


cells and continues beyond 0.01 lumen for cells of lower sensitivity. 
The linearity with voltage for constant light illumination is very good; 
for instance, one very sensitive tube with a grid area of 1 mm wide X 
5 mm long had a drop of only 10 per cent in voltage sensitivity from 
the expected linearity relationship when the voltage changed from 22 
to 90 volts. This was with a light level beyond the linearity region of 
this cell. For lower light levels the linearity with voltage should be 
even better. 


Photoemissive vacuum tubes are flat in frequency response to very 
high frequencies. Photoemissive gas tubes are fairly flat in frequency 
response up to 10,000 cycles but drop appreciably at higher frequen- 
cies. *Lead-salfide cells have a fairly flat frequency response to 10,000 
cycles but then drop quite rapidly above 10,000 cycles. Variations of 
frequency response between individual tubes require additional study. 


It has been observed that sensitivity and frequency response vary 
somewhat at different points of the sensitive surface. This charac- 
teristic of lead-sulfide cells also requires additional study in conjunc- 
tion with the optical systems to be developed for their use. 

Summing up the contents of this paper it seems that the lead-sulfide 
cell is showing considerable advantage in sound reproduction. Its 
application will require a modification of the present mechanical, 
optical, and electrical design of motion picture projectors and the 
optimum operation of these cells will follow an accurate balancing of 
all the factors involved. 


(1) R. J. Cashman, "Lead-sulfide photoconductive cells -for sound reproduc- 
tion," /. Soc. Mot. Pict. Eng., vol. 49, pp. 342-348; October, 1947. 

(2) L. Sosnowski, J. Starkiewicz, and O. Simpson, "PbS photoconductive 
cells," Nature, vol. 159, pp. 818-819; 1947. 

(3) Smithsonian Physical Tables, 1934, p. 386, Table 438. 

(4) A. Cramwinckel, "The sensitivity of various phototubes as a function of 
the color temperature of the light source," J. Soc. Mot. Pict. Eng., vol. 49, pp. 
523-530; December, 1947. 

(5) "V. K. Zworykin and E. D. Wilson, "Photocells and Their Applications," 
John Wiley and Sons, New York, 2nd ed., 1934, p. 187. 

(6) P. Moon, "The Scientific Basis of Illuminating Engineering," McGraw- 
Hill Book Company, New York and London, 1st ed., 1936, p. 146. 

(7) Page 162 of reference* 6. 

Volume Compressors for 
Sound Recording* 



Summary This paper deals in a general way with volume compressors 
of the type used in sound recording. The subject matter is divided into six 
sections: the desirability of volume compression, compressor characteris- 
tics, problems arising from the use of compressors, classification of the types 
of compressors with the advantages and disadvantages of each type, compres- 
sor design, and the measurement of compressor performance. 

MANY DEVICES 1 - 2 have been developed by the communications in- 
dustry for the automatic control of volume in telephonic trans- 
mission. Of these devices, the limited-range compressor and the 
peak limiter have come into general use for 35-mm sound-on-film 
recording, and two others, the volume-operated gain-adjusting de- 
vice 3 and the limited-range expander, may be useful for 16-mm sound 
recording. This paper will be concerned primarily with the problems 
of design and use of compressors and limiters for sound recording 


The range of sound levels to which the ear is sensitive is much 
greater than the range which can be linearly accommodated by any 
known method of sound recording. Fortunately, sound intensities 
which reach the upper threshold of hearing are rare, and very faint 
sounds are usually submerged in the ambient-noise level. Thus, the 
range of sound levels encountered in recording is not greatly in ex- 
cess of the capabilities of 35-mm sound-on-film recording, although 
the range is considerably more than is now practical on 16-mm film. 
Photographic recording is limited at high amplitudes by 100 per cent 
modulation of the exposing light and at low amplitudes by the granu- 
lar structure of the developed image. By manual adjustment of the 
amplification the sound levels to be recorded may be brought to lie 
within these two limits, low levels being brought up and high levels 
suppressed, a process which requires skill and experience on the part 

* Presented May 17, 1948, at the SMPE Convention in Santa Monica. 


50 GRIMWOOD January 

of the operator. Since overmodulation of the exposing light results 
in noticeable and unpleasant distortion, a factor of safety must be 
allowed so that unexpectedly high peak sound levels will not cause 
serious distortion, manual control being too slow to react to sudden 
changes in level. It is evident that if an amplifier could rapidly, and 
without nonlinear distortion, change its gain so that its output level 
were limited to correspond with the overload level of the light modu- 
lator, then the operator could record at a higher average level without 
danger of objectionable distortion of the peak levels. Furthermore, 
if an amplifier could, without distortion, control its own gain so that 
the range of input levels were divided by a factor to result in a lesser 
range of output levels, then a wide range of sound levels could be com- 
pressed into the limited range of the recording medium. If the 
amount of 'such compression were no greater than that normally done 
by manual control, expansion upon reproduction would not be 

Compression of the sound volume range maybe desirable for reasons 
other than the characteristics of the recording medium. For example, 
experimental home recordings on 16-mm film often sound as though 
the volume range reproduced were greater than the volume range of 
the original sound. This effect is probably due partly to the acoustics 
of the rooms in which the recordings were made and partly due to the 
monaural character of the recording process. It is reasonable to ex- 
pect that the naturalness of recordings made in the home (where 
rooms are more reverberant than are sound studios) would be im- 
proved by the use of compressors in recording. 

The volume range that can be reproduced satisfactorily is equally 
as important as the range that can be recorded. The volume range 
that can be used in a motion picture theater is surprisingly narrow, 
much shorter than the range that can be recorded. 4 This is because 
audiences are intolerant of loud sound reproduction and the audience 
noise level is high; some form of compression must, therefore, be 
used in making recordings to be reproduced where the noise level is 
high and the maximum level limited by listener preferences. 


The static input-output relations of a compressor or limiter ampli- 
fier differ from those of an ordinary amplifier in a manner which may 
best be understood by referring to Fig. 1. Curve 0-0' represents the 
input-output relations of a normal amplifier, 0-a-A represents these 




same relations in a compressor amplifier, and O-b-B applies to a lim- 
iter amplifier. Points a and b are known as thresholds. The input- 
output relation of the amplifiers is defined in terms of the slope (deci- 
bel scale) of the curve above the threshold and the input range in 
decibels between the threshold level and the maximum useful out- 
put level. For applications of sound recording, the maximum useful 
output level would be taken as the output level corresponding to 
100 per cent light-modulation. In Fig. 1 an output level of zero 
decibels (point x) has been taken as the output at which the light- 
modulator will be fully modulated. The specification of the com- 




i - 



-20 -<5 -10 -5 5 10 15 20 

input Level -Decibels 

Fig. 1 Idealized input-output characteristics of volume com- 

pressor curve would therefore be: slope l / 2 , range 20 decibels; for 
the limiter curve, the specification is: slope 1 /i , range 20 decibels. 
Alternatively, these two characteristics may be specified more simply 
as a compression of 20 into 10 decibels and as a compression of 20 into 
2 decibels, respectively. When operating with a fixed slope, com- 
pression is sometimes expressed as the decibel difference between 
point x on the linear curve and the corresponding ordinate of the 
compressor curve (#'). This method of expressing the amount of 
compression is not very useful in setting up operating conditions, 
particularly when the slope is very low, as it is for limiter operation, 
but it is very useful in the practical use of a compressor, a meter 

52 GRIMWOOD January 

sometimes being calibrated to give a direct indication of the amount 
of compression. Experimental curves are very similar to those of 
Fig. 1, except that points a and b are not sharply defined. The 
threshold should, then, be denned as the point of intersection of the 
extrapolations of the linear portions of the curve. Since this involves . 
plotting the curves, it is more convenient in practice to define the 
threshold arbitrarily as, for example, the point at which the output is 
compressed by l /z decibel. So long as the method of definition is 
specified, there need be no confusion. 

The values of compression and of limiting used here for illustration 
are representative of those generally used in sound recording. The 
compresssions usually used range from 20 into 10 to 30 into 15, and 
the limiter characteristics are in the region of 10 into 1 to 20 into 1. 

The input-output relations just discussed were defined as static; 
the dynamic input-output relaton must be linear or else the nonlinear 
distortion will be intolerable. The difference between static non- 
linearity and dynamic linearity is one of operating time. If the 
change in gain takes place at a rate so slow that individual cycles of 
the lowest audio frequency to be transmitted are not measurably 
altered in shape, then there will be no nonlinear distortion. Such 
slow operation would defeat the primary purpose of compressors but 
it is also possible to change the gain very rapidly upon application of 
the audio signal and to release the gain to normal very slowly upon 
cessation of the signal. In this case, there will be nonlinear distortion 
only during the short period during which the gain is changing rapidly 
and, if this period can be made sufficiently short, the ear will not de- 
tect the distortion. Compressor amplifiers, therefore, are designed 
to decrease their gain very rapidly upon the sudden application of a 
signal and to increase the gain to normal very slowly upon sudden 
removal of the signal. This timing is controlled by the charge and 
discharge time constants of a simple resistance-capacitance network. 
The time required for a specified percentage completion of gain change 
from the normal (uncompressed) value to the compressed equilibrium 
value upon the instantaneous application of a signal is known as the 
operating time. The time required for the same percentage of gain 
change in the return of the compressed gam to the uncompressed 
equilibrium value upon the instantaneous cessation of the input sig- 
nal is known as the release time. The percentage change in gam for 
which the time is given is usually either 63 per cent (the value re- 
sulting when the time equals the resistance-capacitance product) or 


99 per cent. Both figures are open to objection: the 99 per cent 
figure because it requires an unduly high accuracy of measurement, 
the 63 per cent figure because it gives a misleadingly short operating 
time. The operating time is determined by the charging of a capaci- 
tor through a rectifier and the discharge time by the discharge of the 
capacitor through a fixed resistor. The resistance of a rectifier is a 
function of the voltage drop across the rectifier, increasing as the volt- 
age drop decreases. Thus, when a signal is suddenly applied, the 
capacitor will start to charge according to the exponential charging 
law, but, as the voltage across the capacitor builds up, the voltage 
across the rectifier decreases, the rectifier resistance increases, and the 
capacitor charge builds up to its ultimate value much more slowly 
than would be predicted from the exponential law. An amplitude 
change of 90 per cent of the ultimate change in amplitude is recom- 
mended as a satisfactory compromise in the specification of the action 

The operating time of commercially available compressor ampli- 
fiers is usually from 1 to 2 milliseconds and the release time from 100 
to 400 milliseconds. Experimental evidence 5 indicates that per- 
formance is improved as the operating time is decreased, the lower 
limit being one of practical design. 

The steady-state distortion characteristics of a compressor ampli- 
fier are those of a normal amplifier except at very low audio frequen- 
cies. If a long release time is used, there will be no increase of low- 
frequency distortion, but if a short release time is desired, the timing 
capacitor will discharge sufficiently between cycles of the audio in- 
put to affect the wave form so that some compromise must be made 
between discharge time and distortion. It is possible to devise cir- 
cuits in which the release time would be a function of the duration of 
the applied signal but such refinements have not as yet come into use 
for sound recording. 

Some transient distortions are present in compressors. One of 
these, previously mentioned, is the distortion inherent in changing 
amplification at a rate comparable with the instantaneous rate of 
change of the signal. The gain may, however, be changed in a period 
of time shorter than that required to produce an impression on the 
ear so that this type of distortion is of no practical importance. 
Another, and frequently serious, distortion is the generation of pulses 
known as "thump" from the fast operating time. If a high-frequency 
audio signal is suddenly applied to a compressor amplifier, the static 

54 GRIMWOOD January 

potentials in the variable transmission circuit suddenly change to new 
levels and remain at these levels until the amplitude of the input sig- 
nal is changed. Since the rate of change from one level to the other 
is of the same order of magnitude as the instantaneous rate of change 
of a high-frequency audio signal, the change in potentials will be 
transmitted by the audio amplifier; furthermore, the change from 
one potential level to another will alter the charge on coupling capaci- 
tors between the variable trnsmission circuit and the output audio 
amplifier, and the change in potential on these capacitors as they 
discharge is transmitted by the audio amplifier. Thus, unless some 
means are used to prevent the control signal from appearing in the 
audio output, a sudden increase in compression will result in a pulse 
appearing in the audio output, this pulse taking the form of a sharp 
rise in potential followed by an exponential decay, the duration of the 
pulse being determined by the time constants of the amplifier coupling 

Because of the slow decay time, the audible effect of the pulse is such 
that it is usually called "thump." The term "thump" is often given 
to any audible effect of similar nature; it is not restricted to pulses 
directly due to the control signal. Thump is not necessarily funda- 
mental to compression but is usually present to some degree, just as 
nonlinear distortion is present to some degree in a normal amplifier. 
One of the most serious problems in the design'of a compressor for 16- 
mm recording is the development of a circuit in which the thump com- 
ponent can be held to tolerable levels without frequent inspection and 
maintenance. Thump can be serious out of all proportion to its direct 
audible effect. It may, particularly when operating on a limiter 
characteristic, modulate the audio signal and cause the output ampli- 
tude to exceed that shown by the steady-state characteristic for 
periods of tune long enough for the resulting overload of the light- 
modulator to be plainly audible in the finished record. Tolerable 
levels of thump components will not be given here, but will be dis- 
cussed more fully in one of the following sections. 

The frequency response of a compressor or limiter does not, in 
general, differ from that of a normal amplifier. Because of the com- 
pression action, there are certain requirements for the frequency 
characteristics of the control signal path but they do not necessarily 
affect the audio signal transmission path. The use of a compressor 
does have some effect upon the recording channel as a whole, in that 
the compressor alters and restricts the preferred location of frequency 


equalization in the channel. Both of these effects will be considered 
in some detail in the next section. 

The noise level of a compressor should not differ significantly from 
that of an ordinary amplifier. This means that .the bulk of the system 
noise should arise in sources located ahead of the point at which com- 
pression action takes place. If it arises beyond this point, the signal- 
to-noise ratio will be reduced as the compressor acts. 


The solution of one problem usually gives rise to several others. 
Certainly this is true of the use of compressor amplifiers. One of the 
most serious of these new problems is what has been termed "spec- 
tral-energy distortion." 6 Consider the case of a speech sound which 
starts with a consonant followed by a vowel. Consonants are usually 
composed of high-frequency components of low amplitude, while 
vowels are predominantly low-frequency components of high ampli- 
tude. Therefore, it is probable that the compressor or limiter will 
not be actuated by the opening consonant but will be operated by the 
following vowel. The result (sometimes called "essiness") is over- 
accentuation of some sibilant sounds which, because of long pauses or 
widely fluctuating speech levels, find the compressor in the passive 
condition. If we place in the control path of the compressor an 
equalizer (or "de-esser") which boosts the high frequencies so that 
the compressor responds to lower levels of high-frequency components 
than to those of low-frequency components, then this accentuation of 
sibilants will largely disappear. The effect is entirely eliminated only 
when the compressor threshold fellows a frequency- versus-level curve 
which matches the frequency spectrum of the sound source. 

The spectral-energy distribution of the sound to be recorded is. 
however, a function of acoustic conditions, type of source (speech, 
music, or noise), and frequency response of the recording channel. 
The energy distribution of speech alone varies with the individual 
and with the effort level of the speaker. Thus, the exact compensa- 
tion of spectral-energy distortion becomes a very complex problem. 
Fortunately, satisfactory results are obtained in practice with a single 
"de-esser" equalizer whose characteristics are based upon the average 
spectral-energy distribution of speech. Some further refinement 
probably is desirable in studio recording and can be obtained readily 
by having several equalizer characteristics available to the operator, 

56 GRIMWOOD January 

say, one for each of three effort levels of speech and a fourth for music. 
The frequency characteristic of the "de-esser" equalizer is not, in 
general, the inverse of the spectral-energy distribution of the source, 
but should take into account the average of all frequency discrimina- 
tions, whether acoustical, electromechanical, or electrical in nature, 
between the sound source and the equalizer itself. The correction of 
spectral-energy distortion, while desirable, is less necessary in a lim- 
iter amplifier than in a compressor amplifier because the limiter is 
operated above the threshold less frequently than is the compressor. 
A closely related problem arising from the use of compressor ampli- 
fiers is that of the location of the compressor hi the recording channel. 5 
It is obvious from the input-output relations of a compressor or lim- 
iter that for levels above the threshold the effect of any frequency dis- 
crimination ahead of the compressor will be reduced by the com- 
pression ratio. Since such discrimination is usually intentional, this 
effect is undesirable. Elimination of this effect calls for a rearrange- 
ment of the recording channel such that all equalizing is placed after 
the compressor, but this solution is not entirely satisfactory since the 
mixer operator must have some equalizing under his control in order 
to adjust for the set acoustics and for differences in the source ma- 
terial. Much of the equalizing done by the mixer is of such a nature 
that it automatically corrects for variables that otherwise would dis- 
tort the normal spectral-energy distribution of the source. Thus, a 
satisfactory solution is to place the bulk of the equalizing beyond the 
compressor and to leave a bare minimun of variable equalizing ahead 
of the compressor. If more than one microphone is used simultane- 
ously, the spectral-energy distribution of the sources may be different 
and the equalizing required may be different so that ideally a com- 
pressor should be used in each input circuit. This condition is even 
more likely to exist in re-recording than in original recording. When 
a limiter amplifier is used, it is not permissible to put most of the 
equalizing beyond the limiter. If the limiter is to protect against 
over-modulation of the light-modulator, any equalizing between the 
limiter and the modulator must be restricted to the type which de- 
creases the level of some frequency components. Because the limiter 
acts infrequently and on the highest levels, the effect of equalizing 
ahead of the limiter is not serious. This restriction applies only to 
equalizing in the transmission path beyond the point at which the 
control-signal path branches off from the main path. When the con- 
trol circuit branches from the main path beyond the point at which 


the actual gain-changing takes place, both paths are equally affected 
by equalizing inserted between these two points. Hence, in a limiter, 
the modulator will still be protected against overload and in a com- 
pressor, the equalizing can be considered as being beyond the 

Monitoring the recording level is more of a problem in a channel 
using a compressor or a limiter than in a channel using only normal 
amplifiers. The object of level monitoring is to know, at all times, 
the recording level in terms of the overload point of the light-modu- 
lator. If a volume indicator is placed ahead of the compressor, its 
indication may be correlated with light-modulation at any one fre- 
quency, but, unless the frequency characteristic of the indicator is 
matched to that of the recording channel between the point of connec- 
tion of the indicator and the light-modulator, the indications will be 
of little value! The volume indicator may be placed beyond the com- 
pressor and the equalizer. In this case, the accuracy of reading the 
indicator must be multiplied by the inverse of the compression slope if 
the precision of the indication is to be held to the same value as in a 
channel without a compressor, the input volume range indicated by 
the meter being increased by the amount of the compression. This 
is not necessarily a disadvantage; in fact, a good case can be made for 
the increased volume range shown by the indicator. When a limiter 
is used, a volume indicator beyond the limiter is of little use, since 
a wide range of input levels is compressed into a very small range of 
output levels. If the volume indicator can be given the same f reqency 
response as the limiter, a location ahead of the limiter is satisfactory. 
A meter reading the amount of compression or of limiting may be used 
to supplement the volume indicator, but it is not a satisfactory substi- 
tute because a compression indicator gives no indication of levels 
below the threshold level. 

The release. tuning of a compressor presents some minor problems 
in that the release time should depend partly on the type of material 
being recorded. When the average sound level fluctuates fairly rap- 
idly, as in speech, a short release time is desirable, but when the level 
may change slowly, as in music, a longer release time is preferred. 
Although the release tune can be made to change automatically with 
the duration of the sound, there seems to be little justification for the 
complexity of an automatic control so long as a change in equalization 
(as between speech and music), which is not readily made automatic, 
must be made* by the operator. 





A compressor consists basically of a circuit whose transmission can 
be varied by a control signal and a second circuit which derives this 
control signal from the audio signal. Because the control signal may 
be derived from either the audio input or the audio output, compres- 
sors may be grouped into one of two classes: the forward-acting type 
in which the control signal is derived from the audio input (Fig. 2 (a)), 
and the backward-acting type in which the control signal is derived 
from the audio output (Fig. 2 (b)). These two classes have quite 
different input-output relations; the forward-acting type usually has 
a compression slope which decreases as the input level increases. 
Thus, the input-output curve approaches a maximum and it may have 
a negative slope beyond the maximum. The exact form of the curve 
depends upon the characteristics of the control circuit and of the 


(a) Forward-acting volume 

Fig. 2 


(b) Backward-acting volume 

variable transmission circuit and will change with any shift in the 
characteristics of either of these circuits. 

The compression ratio of the backward-acting circuit is nearly con- 
stant over a wide range of input levels and the input-output curve is 
not greatly changed by the characteristics of the variable transmission 
circuit. The backward-acting compressor has one serious disad- 
vantage which is not present in the forward-acting type. Since it is. 
a form of feedback circuit, self-oscillation is possible and careful cir- 
cuit design is necessary to avoid instability of this type. Compressors 
and limiters at present used in 35-mm sound recording are universally 
of the backward-acting type. Inasmuch as the purpose in using a 
compressor is to reduce the range of signal levels applied to the modu- 
lator without noticeably altering the original volume relations, an 
input-output curve having a constant slope of less than unity (above 
the threshold) is to be preferred to a curve in which the slope de- 
creases gradually from unity to zero, thus completely*destroying the 


syllabic volume relations of high-level input signals. While this 
latter curve is satisfactory for limiter operation, the limiter function 
of preventing overload of the light-modulator makes control of the 
output level by the output level preferable to control by the input 
level. Because of the general acceptance of the backward-acting 
compressor as the more desirable type, the remainder of this paper 
will be devoted to this type, though for the greater part of the text 
it will not be necessary to distinguish between the two types. 

Compressors may be further classified in terms of the form of con- 
trol of the variable transmission circuit. We may term "one-dimen- 
sional" all circuits in which the electrical transmission is controlled 
by an electrical signal. Those compressor circuits in which the 
electrical transmission of the input signal is controlled by another 
form of energy (such as mechanical) will be termed "two-dimen- 
sional." Included in this class are those circuits in which the input sig- 
nal is nonelectrical and the control signal is electrical, ^o commercial 
fast-acting compressors of the two-dimensional type are known to the 
writer, but this type has some very worth-while advantages over the 
purely electrical type if the major problem of slow-action time can be 
overcome, and there are some interesting possibilities. One slow- 
acting compressor of this type uses a thermistor to convert the con- 
trol signal energy into heat which, in turn, controls the transmission 
through the audio signal path. 7 There exists the possibility of com- 
pressing by using the control signal to vary the field strength of the 
magnetic field of a dynamic or of a velocity microphone.* Another 
possibility is a step-by-step compressor which would use high-speed 
relays operated by the control signal to insert attenuators in the 
audio-transmission path.** Another possibility is the adaption of 
the carbon-pile regulator to the high-speed, low-level operation re- 
quired in a compressor. 

These possibilities are sufficient to disclose the two main advantages 
of a two-dimensional system: first, the absence of an electrical con- 
nection between the variable-transmission circuit and the control 
circuit permits the transmission to be altered without the generation 
of transients in the audio path arising from the reaction of the one 
circuit on the other; second, the dynamic linearity of the audio path 
is not affected by the control signal. These two factors are funda- 
mental limitations of the one-dimensional system. If the electrical 

* J. G. Streiffert, of the Kodak Research Laboratories, private communication. 
** T. G. Veal, of the Kodak Research Laboratories, private communication. 

60 GRIMWOOD January 

transmission of one path is to be under the direct control of another 
electrical signal, some special means 'must be employed to prevent the 
control signal from appearing in the first path, and the transmission 
of the desired signal can only be altered by a nonlinear element which 
must simultaneously cause nonlinear distortion of the desired signal. 

The means taken to prevent control-signal components from ap- 
pearing in the audio output may be made the basis of further classi- 
fication of one-dimensional compressors. Three methods have been 
used : the carrier method, the compensation method, and the balance 

The carrier type of compressor 8 uses an oscillator and a balanced 
modulator to shift the audio spectrum up into the carrier-frequency 
range and a demodulator to step back down to audio frequencies. The 
variable-transmission circuit is placed in the carrier link. This cir- 
cuit may be of either the variable-mu or the variable-impedance type, 
although the former ordinarily would be preferred for its simplicity. 
The shift to carrier frequencies makes it possible to separate the con- 
trol signal from the audio signal on a frequency basis and also to 
filter out nonlinear distortion products on a frequency basis. The 
disadvantages of this scheme are that an oscillator, a demodulator, 
and an accurately balanced modulator are necessary. 

The compensator type of compressor 9 - 10 uses a variable-mu type 
of tube to control the audio gain, changes in the plate current of this 
tube due to the control signal being compensated by an opposite 
change in the screen current of a second variable-mu tube. This 
circuit has the advantage of not requiring push-pull operation but 
has the disadvantage of requiring the changes of plate current of one 
tube to be exactly matched by changes in the opposite sense in the 
screen current of another tube. 

Practically all compressors in actual use at the present time separate 
the control signal from the audio signal by balancing the two circuits 
with respect to each other. If the audio signal is applied out-of- 
phase to the two inputs of a push-pull amplifier and the control signal 
is applied in-phase to the same inputs, the control signal anci the audio 
may be separated on a phase basis. This method has two disad- 
vantages: the variable-mu transmission circuit must be push-pull 
(which requires two accurately matched nonlinear elements), and 
some means must be used to cancel the in-phase components appear- 
ing in the output of this circuit. Both requirements can be met 
without great difficulty or circuit complexity. 


Audio-frequency compressor amplifiers may be divided into varia- 
ble-mu types and variable-impedance types. The variable-mu type 
is used almost exclusively for sound-on-film recording and for radio 
broadcasting. Two versions of the variable-mu compressor n ~ 13 are 
currently used : compressors designed primarily for limiter operation 
use mixer-type tubes such as the 6L7, those designed primarily for 
operation over a characteristic curve having a slope of the order of 
one half use remote cutoff pentodes such as the 6K7. The variable- 
mu type of compressor amplifier has one main advantage over other 
types of compressors : simplicity. This circuit simplicity is the result 
of two properties of vacuum tubes; first, the gain of a vacuum tube is 
controllable by the potential of a grid whose impedance is so high 
that for practical purposes there is no power drawn from the source of 
the control signal; second, the vacuum tube can be so used that when 
the amount of compression is a maximum the variable-transmission 
circuit may still have a net gain in signal level, and hence there is no 
net loss in the variable-transmission circuit to be made up by the 
addition of amplifier stages. 

The variable-mu compressor has also a major disadvantage which 
results from the use of tubes to control gain. In order to control the 
circuit gain and at the same time to balance the audio-transmission 
path with respect to the control signal, it is necessary to have two 
matched nonlinear characteristics. The characteristics of vacuum 
tubes, however, vary considerably from tube to tube and drift with 
aging of any individual tube. It is possible, by the use of a negative 
feedback, to make circuits using vacuum tubes linear to any assign- 
able degree. No similar technique is known whereby a circuit may be 
forced to have a given degree of nbnlinearity. Hence, when nearly 
identical nonlinear characteristics are necessary in two vacuum tubes, 
they can be obtained only by a process of selection from a group of 
aged tubes and the matching of the tubes should be checked at fre- 
quent intervals, preferably each time the apparatus is used and at 
least twice during each day of continuous use. This disadvantage of 
the variable-mu compressor is not particularly serious in professional 
sound-recording work because the scale of operations is such that 
routine maintenance and checking of all equipment is standard 

Compressors and limiters of the variable-impedance type comprise 
a very extensive group, not only because a variable impedance may be 
used in many ways to control the transmission of a circuit but also 

62 GRIMWOOD January 

because any nonlinear element has potential applications in com- 
pressor design. Many variable-impedance compressor circuits have 
been published in which the plate impedance of a vacuum tube is the 
variable element. 3>14 ~ 19 Plate impedance may be used in several 
ways : as one portion of a variable voltage divider, as one of the feed- 
back impedances in a feedback amplifier, as a means of effectively 
changing the connections of two transformers from series aiding to 
series opposing; other configurations can be devised. Circuits using 
tubes as variable-impedance elements are subject to the same disad- 
vantage as that cited against the variable-mu circuits. There may be 
some difference in degree since simpler tube types may be used in the 
variable-impedance circuits but matched characteristics are still 
necessary. These circuits have, in general, the characteristic in com- 
mon with other variable-impedance circuits that they are variable-loss 
circuits. This loss must be made up by amplification in some other 
part of the circuit. As in the variable-mu compressors, there is the 
advantage that no appreciable power is drawn from the control-signal 

Some further mention should be made of those circuits in which the 
variable impedance is part of the beta circuit of a feedback amplifier. 
Such circuits are usually accompanied by a claim of superior merit 
because they are feedback circuits, especially in respect to harmonic 
distortion. What merit these circuits possess cannot be attributed to 
negative feedback. As in all other one-dimensional compressor cir- 
cuits, the variable-transmission elements cannot of themselves dis- 
tinguish between the audio signal and the control signal ; both signals 
operate on the same nonlinear characteristic. Distortion is, then, de- 
termined by the curvature of the nonlinear characteristic over the 
maximum range of amplitudes of the audio signal. Strictly speak- 
ing, the usual feedback equations are not applicable to compressor 
circuits; the derivation of these equations assumes linear transfer 

A second group of variable-impedance compressors 20 ' 21 uses 
passive nonlinear elements, known as varistors, in the variable-trans- 
mission circuit. Copper-oxide rectifiers are the most commonly used 
varistors, silicon carbide (Thyrite) has been used in these Laboratories 
in an experimental compressor, and germanium-crystal rectifiers have 
been tried experimentally. The advantage of varistors over vacuum 
tubes is their stability. Two or four units may be selected for matched- 
impedance characteristics and they will remain matched over long 


periods of time. The varistor compressor, in common with other 
variable-impedance types, compresses by introducing a loss of energy 
into the circuit. In fact, the variable-transmission portion of the com- 
pressor is often called the "variolosser." Varistors have one disad- 
vantage not present in vacuum-tube variolossers in that power is re- 
quired to control their impedance. The timing capacitor of the con- . 
trol circuit cannot supply sufficient power without either increasing 
the action time or shortening the discharge time, so that it is necessary 
to insert a direct-coupled impedance-changer tube between the timing 
capacitor and the variolosser. The varistor is a relatively low im- 
pedance device. If it were made sufficiently high in impedance to be 
negligible current drain on the timing condenser, an impractically 
high voltage would be needed for control and the impedance would be 
too high for use in audio-frequency circuits. Also, the discharge time 
would be a function of the amount of compression. 


The purpose of this section is to point out some of the principles to 
be followed in the design of compressor amplifiers; no specific circuits 
will be presented. While much of the discussion will be of general 
applicability, it is intended to apply specifically to those circuits which 
may be represented by the block diagram of Fig. 3. 

The choice of the type of compressor will naturally be arrived at by 
weighing the advantages and disadvantages of the various types in 
relation to the requirements of the particular application. In sound- 
on-film recording, the use of compressor amplifiers may be divided 
among 35-mm and 16-mm apparatus and studio and portable equip- 
ments, the four combinations having different requirements. These 
four uses put different emphasis on such factors as size, weight, cost, 
distortion, flexibility of operation, stability of operating character- 
istics, and routine maintenance requirements. This latter factor is of 
the utmost importance in equipment designed for 16-mm amateur 
sound recording and is probably more important in 16-mm pro- 
fessional work than in 35-mm usage. Cost, size, and weight are all im- 
portant factors in the 16-mm field, even in studio recording, and gains 
can be made in these respects at the expense of flexibility of operation. 
The one factor that cannot be sacrificed, if good 16-mm recordings are 
to be made, is the factor which is usually (and rather vaguely) called 
quality. More exactly, quality means freedom from distortion, per- 
manence of electrical characteristics, and physical durability. 




The desired input-output curve may affect the choice of the type of 
compressor to be used and will greatly influence circuit details. In 
35-mm recording, some studios use a compression of, roughly, 20 into 
10 decibels, while others use a limiting characteristic of about 10 into 
1 decibel. Thus, compressor amplifiers may be designed primarily for 
one or the other of these two types of characteristic. For 16-mm re- 
cording, on the other hand, it now seems probable that the highest 
average quality will be obtained by combining the two characteristics. 
That is, a compression of 20 into 10 decibels should break at a second 
threshold into a compression of 10 into 1 decibel. Thus, the circuit 
will not be the same as would be used for either characteristic alone. 

The choice of a circuit will also be influenced by the operating level. 
In order to reduce thump to a miinmum, it' is necessary to work the 
variable-transmission circuit at the maximum audio level consistent 











L. _ 







Reel - 


1_ . 

Control Circuit 


Fig. 3 Block diagram of backward-acting volume compressor. 

with nonlinear distortion requirements. Hence, if this part of the 
compressor can handle a signal level of 10 volts, to operate it at a 
level of 1 volt would, in effect, increase the thump level by a factor of 10. 
The control circuit, especially that part of it which is associated with 
the rectifier, is very important to the proper functioning of the com- 
pressor. The ratio of discharge to charge time is approximately 1000. 
Thus, if 5 megohms is taken as a practical limit to the value of the 
discharge resistor, the charging impedance should not exceed 5000 
ohms throughout the frequency range to be handled by the com- 
pressor. The rectifier, which for minimum operating time should be 
used in a full- wave circuit, must be fed from a low-impedance source. 
Voltage feedback offers a practical means of obtaining very low source 
impedance ; if the rectifier is fed from a transformer, voltage feedback 
from the primary of the transformer will not be effective unless a high- 
quality transformer is used. The rectifier should be chosen for low 
plate resistance and should be operated with as high a signal level as 


is practical since the plate resistance decreases as the voltage drop 
across the rectifier increases. A high operating level also reduces the 
effect of variations in the contact potential on the threshold level. 
Barrier-layer-type rectifiers do not usually have a sufficiently high 
ratio of back-to-forward resistance to be useful, but the germanium- 
crystal diode may be satisfactory if a sufficient number are used in 

Spectral-energy distortion has been mentioned earlier in this paper. 
Its correction will usually require some boosting of the high fre- 
quencies and perhaps of the low frequencies. If it is desired to do this 
boosting in the rectifier amplifier, it may readily be done in the feed- 
back path. If done in this manner, the amount of feedback remaining 
at the frequency of maximum boost should be enough to hold the in- 
ternal impedance of the amplifier at this frequency to a value suffi- 
ciently low to have little effect on action time. The bulk of the charg- 
ing impedance should be the internal resistance of the rectifier. The 
low-frequency response of the control-signal circuit must be handled 
with care in a backward-acting compressor. Backward-acting cir- 
cuits become unstable when the phase shift around the loop formed by 
the variable-transmission circuit and the control-signal circuit reaches 
90 degrees and the signal level exceeds the threshold level. Phase 
shift introduces a time delay between the point at which the gain is 
increased or decreased and the rectifier so that the gain changes over- 
shoot the correct value on both decreases and increases. The remedy 
for this effect is to keep the phase shift around the loop low, down to 
frequencies below which no signal of amplitude exceeding the thresh- 
old level can reach the gain-changing portion of the circuit. 

It will be apparent from the block diagram of Fig. 3 that the output 
amplifier and the rectifier amplifier can be combined. If the output 
amplifier is made push-pull, the rectifiers may be connected through 
blocking capacitors to the plates of the output stage. A simpler cir- 
cuit is thus obtained though at the expense of flexibility: voltage 
feedback should be used to provide a low source impedance for the 
rectifiers, the power-handling capacity of the amplifier must be suffi- 
cient to supply the peak charging current required by the timing-ca- 
pacitor-charging circuit, the gain of the amplifier must be adequate 
to provide the proper operating level for the rectifiers and must re- 
main fixed, and the frequency response should be such as to give the 
desired degree of compensation for spectral-energy distortion. If the 
amplifier is to supply power directly to a light-valve or recording 

66 GRIMWOOD January 

galvanometer, the only one of these requirements likely to be trouble- 
some is that of frequency response. The equalizing needed to com- 
pensate for 16-mm film losses and that needed to reduce spectral- 
energy distortion are sufficiently similar to make the combining of the 
output and rectifier amplifiers a practical possibility. 

It was mentioned earlier that a combination of compression and 
limiting is desirable for 16-mm recording. This may be done with- 
out very much complication by providing two full-wave rectifiers in 
which the threshold biases and the loop gains of the two rectifier cir- 
cuits have been adjusted to give the desired threshold levels and com- 
pression ratios. There are several means of adjusting the loop 
gains : the compressor rectifier may be fed from cathode followers to 
provide a low-impedance source and the gain set by voltage dividers 
in the grid circuit of the cathode followers; if the limiter rectifier is 
fed from a transformer, the compressor rectifier may be supplied by 
another winding or by taps on the same winding; if a circuit is used 
which requires a direct-current stage following the timing capacitor, 
this stage may be made a double-triode inverter stage so that two 
timing capacitors may be used and the gain set by adjustment of a 
tap on the discharge resistor of the capacitor controlling the com- 
pressor characteristic. In this case, the rectifiers may be fed from the 
same source but must be connected to give control signals of opposite 

In Fig. 3, there is a block labeled "differential circuit." The func- 
tion of this circuit is to remove control-signal pulses before they can 
overload the amplifier. In purely electrical compressor amplifiers, the 
change in gain is always accompanied by a change in the static oper- 
ating point of the variable-transmission circuit, and in the balanced 
type of circuit this change takes the form of in-phase pulses whose 
magnitude may be from ten to one hundred times the amplitude of the 
desired out-of -phase signal. With a bridge configuration of the vari- 
able-transmission circuit controlled by a truly push-pull control sig- 
nal, these pulses would not appear at the audio terminals of the bridge 
but, except in the doubly balanced type, they do appear in the audio 
circuits and usually of such amplitude that they will overload the 
output amplifier unless removed by a differential circuit. Perhaps the 
best-known differential circuits are the transformer and the push- 
pull choke. An in-phase signal applied to either produces no net flux 
so that a push-pull transformer or choke is effectively a short circuit 
for in-phase components of the signal. Differential response may be 


obtained with vacuum-tube circuits. 22 Three variants of a basic 
differential tube circuit have been used in these Laboratories in this 
application. By using some negative feedback these circuits are 
easily made so stable that after the initial adjustment their charac- 
teristics remain unchanged by tube aging or replacement. Either 
single-sided or push-pull output may be used and moderate amplifi- 
cation can be realized. 

Because the static operating point of the variable-transmission 
circuit changes with the amount of compression, reactances asso- 
ciated with this circuit may be a source of trouble. For example, the 
plate and screen currents of variable-mu tubes change with the 
amount of compression. These circuits should be returned directly 
to a plate supply of good regulation, not decoupled by a resistance- 
capacitance filter. If a resistance-capacitance decoupling circuit is 
used, it may be responsible for either an overshooting or a slow 
creep of the compressed signal, depending upon the time constant 
of the resistance-capacitance circuit. 

Although variable-mu tubes present a serious problem in the 
matching and maintenance of their characteristics, they are widely 
used for the variable-transmission circuit of commercially available 
compressors. The tube types most used are the 6L7 with the audio 
input to No. 1 grid and the control signal to grids No. 1 and No. 3, 
and the 6K7 with both signals applied to No. 1 grid. The stability of 
the operating characteristic may be improved by a very large resistor 
common to the cathodes of the push-pull variable-mu tubes. If an- 
other tube is used as a cathode resistor, a high impedance is obtained 
without an excessive voltage drop, and by applying the control signal 
to the grid of this tube, the sensitivity is greatly increased. The 
high common cathode impedance also makes the variable-mu stage 
self -in verting so that a push-pull input stage is not essential. With 
this type of compressor, a balance control is needed to adjust the cir- 
cuit for minimum thump from time to time. This control may be a 
potentiometer of roughly 200 ohms connected between the cathodes 
of the variable-mu tubes, with the slider wired to the plate of the 
control tube. A triode-connected variable-mu pentode seems to be 
the best choice for the control tube, the internal impedance of the tube 
and the plate-operating voltage of the compressor tubes being deter- 
mined largely by the choice of cathode resistor for the control tube. 
The combination of a push-pull stage with a large common cathode 
resistor results in nearly complete cancellation of even-order 

68 GRIMWOOD January 


harmonic distortion but, at high audio levels, a form of thump may 
be present which can be traced, not to compression action, but to 
the changes in average plate current which accompany even-order 
distortion, these changes being not necessarily best balanced when 
the balance for compression thump is optimum. 

Both Thyrite and copper-oxide varistors have been used in these 
Laboratories in experimental compressors of the variable-impedance 
type. Thyrite is a high-resistivity material in which the current is 
proportional to a power of the voltage, the power being about 4 for 
disks made to operate in the range of 1 to 20 volts across the disk. 
The nature of the material limits the use of Thyrite to relatively high- 
level operation, the smallest disks made by the General Electric Com- 
pany handling an audio level of 1 volt with less than 1 per cent dis- 
tortion. The characteristics of Thyrite are satisfactorily stable, the 
main problem in its use in a compressor being that of obtaining pre- 
cisely matched disks. This problem can be overcome by selecting 
disks from a large group or, perhaps more economically, by cutting 
the disks from a rod form and grinding the pieces to the exact size. 

Copper-oxide rectifier disks of the type used for modulators are 
very satisfactory for use in compressors and are obtainable in matched 
groups, matching of the disks by a process of selection being accepted 
practice in making balanced modulators. A single copper-oxide disk 
of this type will handle between 10 and 15 millivolts of audio signal 
with not more than 1 per cent distortion. By using a number of 
disks in series, higher levels may be handled. 

Either type of varistor may be used as one arm of a voltage di- 
vider in the grid circuit of a vacuum tube. The varistor should, 
however, be used as the shunt arm since this portion of the divider 
has across it the lower audio voltage. From a circuit point of view, 
it is more convenient to use four varistors in a bridge network, each 
half of the bridge being the shunt portion of a voltage divider. Since 
the varistor network is carrying direct current, it will ordinarily be 
isolated from the preceding amplifier and the following differential 
amplifier by coupling capacitors. If the varistors are not well 
matched, the charging current of these capacitors will exaggerate the 
thump, so a thump level which is not audible may cause circuit diffi- 
culties. Similarly, the circuit preceding the varistor network should 
be of high impedance so that the two sides of the push-pull input cir- 
cuit need not be well matched. If the input circuit is a phase in- 
verter, it should not be the phase-splitter type which has low internal 


impedance for one phase and high impedance for the other. As a 
matter of more theoretical than practical interest, it might be men- 
tioned that if the source impedance is sufficiently high, the varistor 
network can be fed from a single-sided source without increasing 
thump, this type of feed being impractical because there is no longer 
cancellation of even-order distortion. 

When the amount of compression is large, the audio level across the 
varistor network will be high at the instant of application of the sig- 
nal. For example, if there is a 20-decibel compression, the audio sig- 
nal appears across the varistor at ten times its ultimate amplitude, 
tinder these conditions, there are likely to be thump pulses appearing 
in the output which are not consistently repeatable either in ampli- 
tude or in polarity and the action of the compressor in response to a 
suddenly applied audio input will, when examined with an oscilloscope, 
at times appear perfect and at other times will show evidence of se- 
vere overcompression. This effect may be reduced to negligible pro- 
portions by using biased diodes across the varistor network to limit the 
instantaneous amplitude to a value slightly above the steady-state 
level. Another method, which unfortunately would add consider- 
ably to the circuit complexity, is to combine some of the advantages 
of the forward-acting compressor with those of the backward-acting 
type. It is practical in a forward-acting compressor to delay the 
audio signal sufficiently to prevent the instantaneous audio level 
from ever becoming excessively high by inserting a delay network in 
the audio path beyond the point of connection of the control-signal 
path. 17 It should then be possible to feed the output of the control- 
signal rectifier into the same timing circuit as is used for the backward- 
acting portion of the compressor and, by proper adjustment of gain 
and threshold, to obtain a condition in which the instantaneous audio 
amplitude is limited by the forward-acting circuit while the equilib- 
rium input-output characteristic is determined by the backward-act- 
ing circuit. 

Because varistors change their impedance in response to the current 
through them, a direct-current amplifier stage is necessary between 
the timing capacitor, which cannot supply appreciable current, and 
the varistor network. Of a number of circuits which might be used 
in this position, the one, known as the Voltohmyst circuit, seems to 
have the most advantages. The transconductance of the circuit can 
be controlled by the relative values of the individual and the common 
cathode resistors, the use of fairly large individual cathode resistors 

70 GRIMWOOD January 

makes the circuit adequately independent of tube changes, and there 
is partial phase inversion by virtue of the common cathode resistor. 
This latter point is of importance because the in-phase signal at the 
audio terminals of the varistor bridge decreases as the phase inversion 
of the control becomes more perfect. The circuit also has the advan- 
tages that it is adaptable to control signals of either polarity and the 
normally unused grid is available for use with auxiliary circuits. It 
is desirable for two reasons to operate the control stage in a slightly 
unbalanced condition: a very slight shift from a perfectly balanced 
condition may result in a reversal in the direction of current flow 
through the varistors at the start of compression, which will have a' 
serious effect on the compression action. Also, the varistors should 
be biased to give an insertion loss of about 6 decibels, or their imped- 
ance will have to change greatly before they cap cause any appreci- 
able compression. 


Thorough testing of audio amplifiers requires a formidable array 
of equipment, such as an audio oscillator, a calibrated attenuator, an 
audio-frequency voltmeter, a cathode-ray oscilloscope, distortion- 
measuring equipment, and a square-wave generator. All these 
items are useful in measuring the performance of compressor ampli- 
fiers, and additional specialized equipment is desirable for measure- 
ments of timing and thump. 

The steady-state characteristics of a compressor amplifier may be 
measured with the equipment and techniques ordinarily used in test- 
ing audio amplifiers and need not be discussed here except in so far as 
the tests are modified by the peculiarities of compressors. The input- 
output relation should be measured and plotted on decibel scales. 
While such a graph is needed for a complete specification of the input- 
output relation, most backward-acting compressors may be specified 
satisfactorily in this respect by numerical values for the threshold 
level and the range and slope of the compression region. The range 
and slope are adequately specified by expressing the useful working 
range above the threshold in decibel-input range and decibel-output 
range. For example, 20:10 decibels means that the output level 
increases 10 decibels above the threshold level for a 20-decibel in- 
crease in input level. The threshold level may be defined as* the out- 
put level which is 1 /2 decibel below the uncompressed output level. 
This definition of threshold level is recommended because most 


compressor amplifiers have provision for turning off the compres- 
sion. Hence, under actual operating conditions, this definition al- 
lows the threshold level checked very easily. 

Frequency response, noise level, and distortion should be measured 
in both the compressed and the uncompressed condition and should 
be specified for the maximum operating output level in the com- 
pressed state. Noise level and distortion should also be given for 
the same output level in the uncompressed condition, and frequency 
response should preferably be given for both low- and high-output 
levels. Distortion should be measured at a number of frequencies, 
particularly in the compressed state and at low frequencies. A com- 
pressor amplifier should also be checked for instability by making 
a frequency run at high-input level covering the lower audio range 
down to about 5 cycles per second. Instability is most likely to 
occur in the frequency range of 10 to 20 cycles per second and then 
only at input levels high enough to cause compression. 

The tests so far described differ from those ordinarily used only in 
extent and thoroughness and, while they adequately specify the 
steady-state performance of a compressor, they give no information 
as to what happens during and immediately following a change in 
gain. Meeting steady-state performance requirements is not diffi- 
cult. It is the manner in which a compressor makes the transition 
from the uncompressed to the compressed condition, or vice versa, 
that will usually determine the merit of a compressor amplifier. The 
transition performance of a compressor can be specified fairly well from 
measurements of the action times and the thump level. Action 
times were defined earlier, and it was pointed out that there is no 
generally accepted definition. Consequently, a figure for operating 
time or release time is meaningless unless accompanied by a definition 
of these times. For the reasons given earlier, the writer prefers to 
define action time as the time interval between the start of compres- 
sion or decompression (in response to an instantaneous change of 
level) and 90 per cent completion of the change in gain. Both times 
should be specified for a stated amount of compression because the 
change in gain is not necessarily linear with the voltage on the timing 
capacitor, so that the rate of change of gain will depend not only on 
the rate of change of the timing-capacitor voltage but also on the 
voltage level, and in addition, the charging rate of the capacitor is 
not exponential because the rectifier impedance is a function of the 
charging current. 

72 GRIMWOOD January 

The measurement of action times is difficult and requires equipment 
which is not commonly available. The times have been defined in 
terms of the amplifier gain and gain can b$ measured only when there 
is an input signal. The measurement of action times therefore re- 
solves itself into a measurement of the growth and decay times of the 
envelope of an audio-frequency signal. The release time may be 
measured by switching a high audio-frequency input signal from a 
level which gives a chosen amount of compression to a lower level 
which gives an output below the threshold level. An oscillogram of 
the output signal can then be measured to determine the release time. 
A variable-width sound recorder can be used as a recording oscillo- 
graph but is not satisfactory for the measurement of release time be- 
cause the narrow track width and high rate of film travel gives a trace 
which has a small amplitude and a long base line. A cathode-ray 
oscilloscope having a long persistence screen, a single sweep, and cali- 
brated tune markers could probably be used with fair accuracy with- 
out the necessity of making a photographic record. 

The measurement of operating tune is more difficult than the meas- 
urement of release time because the action is so short, preferably less 
than 1 millisecond, and hence the operating time becomes comparable 
to the period of the highest audio frequency which can be transmitted 
by the amplifier. Also, the presence of thump tends to obscure the 
amplitude changes due to the operating time. Fortunately, the actual 
value of the operating time becomes less and less important as the 
time becomes shorter and shorter; the object of a very short operating 
time is largely to prevent overload of the light-modulator. Conse- 
quently, visual examination of recordings made on a variable-width 
sound recorder of the highest audio frequency that can be recorded 
will give all the information regarding operating time that is really 

If a pulse generator and a single-sweep oscilloscope (preferably 
with a long persistence screen) are available, the operating time may 
be estimated with fair precision without the delay involved in making 
a recording. When a single pulse of known duration, say, 1 milli- 
second, is applied to the input of a compressor, the pulse amplitude 
being set to equal the peak amplitude of the input level for which the 
operating time is to be determined, the pulse appearing at the output 
of the compressor will have a sharp leading edge of high amplitude 
which decays exponentially into a flat top, the flat top persisting for 
the remainder of the duration of the pulse. The fraction of the pulse 


width at which the exponential portion of the output pulse has dropped 
to 10 per cent of its initial amplitude (the flat portion being taken as 
zero) is the operating time in milliseconds. 

An oscilloscope is invaluable for examining the action of a compres- 
sor for defects. If a signal of 5000 cycles per second is suddenly ap- 
plied to the compressor-input terminals, the output signal as viewed 
on the screen of a scope with the sweep set to a low frequency, say, 
20 cycles per second, should show a pattern of constant amplitude 
except for a pulse of duration of 1 cycle or less of the input signal, 
this pulse appearing at the leading edge of the pattern. At such low 
sweep rates, the individual cycles of the pattern will not be resolved 
but the envelope of the output signal will be clearly visible. If the 
pattern shows a rounded envelope joining the first pulse to the flat 
portion of the pattern, indicating that a number of cycles have am- 
plitudes exceeding the final amplitude, the charging time is too long, 
probably because the rectifier has too high an internal impedance or 
is being operated at too low a voltage, so that the timing capacitor is 
partially charged very quickly but the remainder of the charge is ac- 
cumulated slowly through the increasing impedance of the rectifier. 
When the input switch is closed, the output pattern should appear 
instantly to the eye and remain steady without bounce or creep, 
these effects being most probably due to storage of energy by a react- 
ance in some part of the circuit where the current drain is a function 
of the amount of compression. Bounce is likely to be caused by a 
storage circuit of short time constant and creep by a circuit of long 
time constant. 

Since thump is the visible evidence of the chief problem in com- 
pressor design, methods by which it may be measured and specified 
are of great importance. Routine checking of compressor perform- 
ance is sometimes done by measuring the degree of in-phase balance. 
The in-phase balance check is made simply and quickly without the 
use of special equipment but has two disadvantages: First, the test, 
as used, is specific for each model of compressor amplifier and is not 
easily generalized to be equally useful on any and all compressors. 
Second, optimum in-phase balance and minimum thump do not 
necessarily represent identical operating conditions for all compressor 
designs. A method of directly measuring thump, which has been 
used by one of the motion picture studios, is illustrated 12 in Fig. 4. 
A 7000-cycle carrier is switched on and off by the 2-cycle oscillator and 
an electronic switch, switching disturbances are filtered out by a high- 




pass filter, and the remaining signal is applied to the compressor to be 
tested. The 7000-cycle carrier is removed from the compressor out- 
put by a low-pass filter, leaving the low-frequency thump compo- 
nents which are measured with a volume indicator. The thump 
components should measure at least 55 decibels below the unfiltered 
output. This method is applicable to the measurement of any type 
or design of compressor and it measures thump relative to the signal 
amplitude. For experimental work on compressors, it is best to 
make some modifications in this method of testing. The switching 
rate should be variable and the volume indicator should be replaced 
by a cathode-ray oscilloscope. The switching rate should be adjust- 
able because thump sometimes is a maximum at a particular rate, 


| Compressor ' 
"I Under Test I" 
I I 

I 1 

Fig. 4 Block diagram of circuit for testing volume- 
compressor action. 

and the lowest rate should allow the period during which the signal 
is off to be about twice the release time of the compressor in order to 
allow a return to normal gain during the off period. The highest rate 
need not be greater than ten interruptions a second. A scope is rec- 
ommended in place of a meter because the indication of a meter for 
pulses of short duration depends upon such factors as the duration 
and repetition rate of the pulses and the dynamic response of the 
meter, and even though the dynamic characteristics of the meter were 
standardized (as they are for the volume indicator), the meter reading 
and the permissible thump level would have to be correlated for every 
interruption rate to be used in testing. The use of an oscilloscope 
permits the actual peak amplitude of thump components to be meas- 
ured relative to the peak amplitude of the signal and also allows quali- 
tative examination of the wave form of the thump. 

There remains the question as to what level of thump is permissible 
and to this question no entirely satisfactory answer is known to the 
writer. Listening tests made on direct-speech-pickup and music 
reproduced from vertical-cut transcriptions have indicated that a 
thump level (measured with a scope as described in the preceding 


paragraph) of 5 per cent is not perceptible. These tests were made with 
a compression slope of one half and the amount of compression 
reached about 15 decibels on signal peaks. The results should not be 
taken as conclusive; operation on a slope of Vio (limiter) may re- 
quire a lower thump level and thump may be more serious in sound 
recording than it is in direct monitoring. A thump level of 5 per cent 
is rather high, compressor operation is fairly satisfactory, but limiter 
operation with this level of thump is not. The thump adds to the 
signal amplitude so that the output level, instead of being controlled 
by the signal amplitude, is controlled by the amplitude of signal plus 
thump. When the loop gain is high, overcompression may result and 
recovery from this condition is delayed by the release time. In work- 
ing with compressors using the varistor bridge, it has been found that 
when thump was reduced from 5 per cent to 1 or 2 per cent the balance 
and stability of the circuits associated with the varistor circuit became 
much less critical factors. Perhaps the most satisfying answer to 
the question of permissible thump level is that if the operation of a 
compressor amplifier, when examined by the interrupted-signal 
method, appears clean and consistent under all operating conditions, 
then the thump level is unimportant. 


The need for compressors in sound-on-film recording and the prob- 
lems inherent in their use have been discussed, followed by an analy- 
sis of compressor characteristics, the types of ^compressors available, 
and the methods of evaluating their performance. Although im- 
provements and refinements can be expected, the volume compressor 
has reached a state of technical development which makes it a neces- 
sary part of sound-recording equipment. The technique of measur- 
ing and specifying performance has not kept pace with this technical 
development. There is a need for standardization of nomenclature 
and test procedure that will permit exact specification of volume- 
compressor performance. 


(1) S. B. Wright, "Amplitude range control," Bell Sys. Tech. /., vol. 17, pp. 
520-538; October, 1938. 

(2) A. C. Norwine, "Devices for controlling amplitude characteristics of tele- 
phonic signals," Bell Sys. Tech. J., vol. 17, pp. 539-554; October, 1938. 

(3) S. B. Wright, S. Doba, and A. C. Dickieson, "A Vogad for radiotelephone 
circuits," Proc. I.R.E., vol. 27, pp. 254-258; April, 1939. 


(4) W. A. Mueller, "Audience noise as a limitation to the permissible volume 
range of dialogue in sound motion pictures," J. Soc. Mot. Pict. Eng., vol. 35, pp. 
48-59; July, 1940. 

(5) M. Rettinger and K. Singer, "Factors governing the frequency response 
of a variable-area recording channel," J. Soc. Mot. Pict. Eng., vol. 47, pp. 299-327; 
October, 1946. 

(6) B. F. Miller, "Elimination of spectral-energy distortion in electronic com- 
pressors," J. Soc. Mot. Pict. Eng., vol. 39, pp. 317-324; November, 1942. 

(7) J. A. Becker, C. B. Green, and G. L. Pearson, "Properties and uses of 
thermistors thermally sensitive resistors," Bell Sys. Tech. J., vol. 26, pp. 170- 
212; January, 1947. 

(8) U. S. Patent No. 2,379,484, Robert L. Haynes, assigned to RCA (1946). 

(9) W. H. Stevens, "Variable slope with constant current," Wireless Eng. 
(London), vol. 21, pp. 10-12; January, 1944; Electronic Ind., vol. 3, p. 176; March, 

(10) A. N. Butz, Jr., "Surgeless volume expander," Electronics, vol. 19, pp. 
140-142; September, 1946. 

(11) W. L. Black and N. C. Norman, "Program-operated level-governing 
amplifier," Proc. I.R.E., vol. 29, pp.. 573-578; November, 1941. 

(12) J. K. Milliard, "The variable-density film-recording system used at MGM 
studios," J. Soc. Mot. Pict. Eng., vol. 40, pp. 143-176; March, 1943. 

(13) J. P. Taylor, "Limiting amplifiers," Communications, vol. 17, pp. 7-10, 
39-40; December, 1937. 

(14) R. C. Mathes and S. B. Wright "The Compandor an aid against static in 
radio telephony," Bell Sys. Tech. J., vol. 13, pp. 315-332; July, 1934. 

(15) G. W. Cowley, "Volume limiter circuits," Bell Labs. Rec., vol. 15, pp. 311- 
315; June, 1937. 

(16) G. Q. Herrick, "Volume compressor for radio stations," Electronics, vol. 
16, pp. 135 and 323; December, 1943. 

(17) D. E. Maxwell, "CBS automatic gain-adjusting amplifier," Tele-Tech, 
vol. 6, pp. 34-36, 128; February, 1947. 

(18) L. B. Hallman, Jr., "Practical volume compression," Electronics, vol. 9, 
pp. 15-17, 42; June, 1936. 

(19) H. H. Stewart and H. S. Pollock, "Compression with feedback," Electron- 
ics, vol. 13, pp. 19-21; February, 1940. 

(20) S. Doba, Jr., "Higher volumes without overloading," Bell Labs. Rec., vol. 
16, pp. 174-178; January, 1938. 

(21) O. M. Hovgaard, "A volume-limiting amplifier," Bell Labs. Rec., vol. 16, 
pp. 179-184; January, 1938. 

(22) J. F. Toennies, "Differential amplifier," Rev. Sci. Instr., vol. 9, pp. 95- 
97; March, 1938. 

Some Distinctive Properties of 
Magnetic- Recording Media* 



Summary Information is presented relative to the adjustment of bias 
current in magnetic recordings and the various effects of bias changes on 
distortion, frequency response, overload characteristics, and permanency 
are discussed. Other factors which influence frequency response are out- 
lined briefly and it is shown that the inherent frequency response of the 
medium is difficult to divorce from effects due to the recording system. 
The problem of noise is presented in general terms and the nature and level 
of the noise from a direct-current saturated medium is advanced as an im- 
portant criterion of quality. 


IN THE COURSE OF the authors' research in connection with the devel- 
opment of magnetic-recording tapes, it has become necessary to 
develop techniques for rapid evaluation of tapes of very widely 
different properties. These techniques are exactly those required by 
a user who wishes to get the best possible performance from a mag- 
netic tape, and thus are of some general interest. The remarks are, 
for the most part, quite generally true of such other magnetic media 
as wires, sheets, disks, and cylinders. Many of the important char- 
acteristics of a magnetic material may be best obtained by measure- 
ments of its magnetic properties, but only data of the sort obtained by 
conventional erase, record, and reproduce heads are discussed here. 
The data were taken on V^hich magnetic tape on a loop tester em- 
ploying a modified Ranger erasing head, a modified Brush recording 
head, and an unmodified but selected Brush reproducing head. 
Various speeds have been used from 5 to 36 inches per second; most 
data were taken at 9.2 inches per second. 


It has been recognized for several years that to obtain good repro- 
duction from magnetic media it is desirable to add to the audio cur- 
rent in the recording head a certain amount of high-frequency bias 

* Presented May 18, 1948, at the SlkPE Convention in Santa Monica. 





current. 1 From the point of view of distortion, the frequency of the 
bias current is not critical so long as it is high enough not to beat with 
any appreciable harmonics of audio current; also, it will be shown 
later that from the point of view of noise it is desirable to have the 
bias frequency high. However, the current value must be selected 
with care. It would seem simple enough to vary the bias current 
until best results occurred, but evidently it is easy to err since many 
conflicting systems have been used for setting bias. Yet the bias 
value is most important in obtaining maximum output with minimum 




. ., . 


Fig. 1 Effect of bias current on output obtainable with 1 per cent 
third-harmonic distortion, at 400 cycles per second. 

distortion, and we believe that it can be selected systematically, at 
least for media of desirable characteristics. 

Most frequently one sees the effect of bias current depicted by 
curves showing how the output (playback) level of a tape signal and 
the distortion of this signal vary with bias current for a constant audio 
recording current. These curves are easy to obtain experimentally, 
but difficult to interpret. However, one is not usually concerned 
with the input current required. The curves of Fig. 1 were plotted 
with this in mind. Here the input is not held constant; the curves 
show the maximum level obtainable at 400 cycles per second with 1 
per cent third-harmonic distortion, as a function of bias current. We 
measure third-harmonic distortion in preference to total distortion 




since the latter may be affected by noise level, whereas the former 
may be related theoretically with the magnetic properties of the 
medium. Even harmonic distortion is negligible in magnetic record- 
ing if the recording is done on a magnetically neutral tape by alter- 
nating-current bias of good wave form. Even harmonic distortion 
will result from magnetized heads or the use of a direct-current (per- 
manent-magnet) erase. The curves are for three " types" of oxide 

80 90 


Fig. 2 -Variation in distortion with output level, 
near the overload point, 400 cycles per second. 

tapes into which we find the hundreds of oxides tested may be divided 
roughly, although no two will be exactly alike. Unless other con- 
siderations may be shown to enter, we would start by choosing the 
bias to give a maximum in such a curve as shown in Fig. 1. 

Regarding the three shapes of the curves, one may say that it is 
preferable to have a broad maximum rather than a sharp one simply 
from a design and control point of view, but it is also desirable quite 
generally. The sharp maxima are usually associated with bad over- 
load characteristics and also with an output sharply varying with 




small bias changes for constant input. This will apply to sharp 
maxima as in curve (C) or the first sharp maximum in the curve (A). 
For example, Fig. 2 shows some overload characteristics where third- 
harmonic distortion is plotted on a log scale against output in decibels 
and it may be seen that the sharp maximum in curve (C) of Fig. 1 
which corresponds to the irregular overload characteristic of Fig. 2 is 
not a desirable bias setting, and the bias should be increased perhaps 
20 per cent to a value which gives a smooth overload characteristic. 
On the other hand, overload characteristics are good for bias set for 
the maximum of curves (A) and (B) in Fig. 1; the data for tape A 




I I 

I I 

2. .512 5 10 



Fig. 3 Frequency response for constant-current recording at 9.2 
inches per second, as affected by bias current. Brush heads. 

are shown in Fig. 2 In general, if the bias setting seems very critical, 
it is an unsatisfactory tape or an incorrect bias setting. 

The discussion above is based on data at one frequency and there 
may be some question about whether the best bias for one frequency 
is the best for another. There has been a tendency to standardize 
tests at 400 cycles per second which we believe should be continued 
until there is evidence for some better test frequency or frequencies. 
One-definite result noted with all media is that a higher bias current 
leads to poorer high-frequency response. It is believed that this 
effect is caused by the partial erasing action of stray bias field and thus 
is somewhat a function of the design of the recording head, but wha't- 
ever the cause it leads one to use the lowest value of bias possible if 




the maximum in Fig. 1 is so broad as to allow a variation in bias with- 
out sacrifice of signal level. For example, Fig. 3 shows two unequal- 
ized frequency-response curves (constant-current recording) made on 
the tape of Fig. 1, curve (A), one with a high (0.6-ampere) and the 
other with a low (0.13-ampere) bias. These values of current are 
those used in a separate 20-turn bias winding on the Brush recording 
head. It can be seen that this increase of bias leads to a very large 
loss of highs. Ordinarily for high-quality work, the bias will be set 
without regard to frequency response and the speed or equalization 
adjusted to give the required high-frequency, response. 






.1 2 3 .4 J5 .6 .7 


Fig. 4 Stability of recordings in small alternating-current 
fields and erase currents required to obliterate, as affected by 
bias current used in recording. 

There are other slight frequency dependent effects of bias which are 
not well understood, but which are relatively minor for a good head 
and tape design. These are connected with the nonuniformity of the 
bias field through the thickness of the medium at the recording head 
and the relative effectiveness in playback of surface and subsurface 
layers of the magnetic medium as a function of recorded wavelength 
and probably with other factors which we do not understand. 

Another effect of bias current which is of interest and possible 
importance is its effect on ease of erase. Fig. 1 shows that it is pos- 
sible to record with two different bias currents and correspondingly 
different audio currents and get the same level recording with the same 


distortion. Two such recordings will not differ in playback, but are 
very different in permanency as Fig. 4 shows. In this figure is 
plotted the remaining level of signal as successively higher erase cur- 
rents are used in the erase head. It is seen that the signal recorded 
with high bias is harder to erase and much less affected by weak 
alternating-current fields, which may be an advantage or disadvan- 
tage depending upon the use to be made of the recording. 


The subject of frequency response of a medium is one which has been 
treated rather thoroughly in previous papers in most respects. It is 
impossible to specify the frequency response of the medium as such, 
since, at least at the present state of the art, it is always affected to 
some degree by the heads used in the recorder. The effect of gap 
width in the playback head is discussed by Holmes and Clark, 1 and 
of gap design in the record head by Clark and Merrill. 2 The bias 
used affects the frequency response probably in accordance with the 
record-head design. The problem of azimuthal alignment of record 
and playback heads always enters in some slight degree for the 
shortest wavelengths that may be recorded. Demagnetization in 
a tape is probably an important factor, but not so important as it 
was once thought to be 3 or as it apparently is in wire recording. The 
part which coercive force plays in the effect of demagnetization may 
be considerable, and thus gives a large variation in frequency re- 
sponse so far as the medium is concerned, but we have found that 
this effect may be negligible in practice. The reason is that with the 
higher coercive force a larger bias field is normally required, and the 
effect of the larger bias field in reducing high-frequency response may 
be of the same order as the increase in high-frequency response re- 
sulting from the higher coercive force. Thus with our particular 
recording head and operating with bias adjusted as discussed above, 
we have found the frequency response not to be simply related to 
coercive force. Conceivably, with a more ideal recording head, a 
direct relationship could be observed, and with a very poor head an 
inverse relationship is possible. In any event, we have come to re- 
gard frequency response as a less important characteristic of a tape 
than we once did. It appears that in an economical "home" recorder, 
where it is desired to get optimum performance at a low tape speed 
(7.5 inches per second), the quality is more limited by distortion, 
hum, and inexpensive components than by frequency response, while 




in a high-quality machine, operating at 18 or 30 inches per second 
tape speed, the frequency response is usually adequate with any of a 
wide variety of tapes. 

Two factors affecting the frequency response of a tape other than 
its magnetic properties, are its thickness and its smoothness. Re- 
garding thickness, it appears that the signal from a given tape at very 
short wavelengths is obtained almost solely from a very thin surface 
layer, so that changes in thickness above a very thin minimum do 
not affect the signal at high frequencies. On the other hand, a signal 


I I I I 






[001 IN) 

Fig. 5 Attenuation in signal strength caused by various separa- 
tions between recorded tape and playback head, as a function of 
frequency (at 9.2 inches per second tape speed) or wavelength 
(at any speed). 

of long wavelength utilizes the full thickness of the medium up to any 
practicable dimension, and thus is affected by changes in thickness. 
For a given recording current and bias, the signal obtained will be 
somewhat (though not directly) proportional to thickness at low 
frequencies, and roughly independent of thickness at high frequencies. 
One qualifying restriction, however, to an unlimited increase in 
thickness for better bass response comes, once again, from head 
design. For the usual ring-type recording head, the fields, both audio 
and bias, drop off rapidly with distance from the head, and there is a 
limit to the tape thickness which may be used without suffering 
distortion from the gradient existing in the bias field. To this extent, 


the optimum tape thickness is highly dependent upon the nature of 
the field of the recording head. 

The surface of the tape (and the heads) must be as smooth as 
possible for good high-frequency response. From simple dipole 
theory, one may expect the field from a very short wavelength record 
to fall off very rapidly with distance from the tape, and this is indeed 
the case. Fig. 5 shows the results of an experiment in which various 
frequencies were recorded on a tape which was then played back with 
the playback head in contact and also separated from the tape by 
various thin paper shims. The playback signal level relative to that 
obtained in contact is plotted against frequency so that the attenua- 
tion introduced by various separations is shown as a function of fre- 
quency (or wavelength) . The dotted curve is an extrapolation to a 
very small separation of 0.0001 inch (2.5 microns) and it can be seen 
that even this minute separation causes a noticeable attenuation at 
very short wavelengths. Thus, while it may be practicable so far as 
equalization is concerned to record wavelengths of the order of 0.001 
inch, more reliable and reproducible results will be obtained if higher 
speeds are utilized to keep the minimum recorded wavelength longer. 
The use of a longer minimum wavelength also simplifies other prot*- 
lems such as head azimuth adjustment. 


The subject of noise is the one in which there is probably least 
agreement among various investigators. A treatment of erased noise 
and noise as a function of direct-current magnetization is given by 
Holmes, 4 based on measurements on wire. The noise on a tape may 
theoretically be reduced to that produced by randomly oriented 
domains by an erasing head which truly demagnetizes the tape, and 
probably the dominant noise such a tape should produce is the slight 
microphonic noise in the playback head due to tape friction. Our 
experiments show that this condition may be approached, based on 
tapes demagnetized by relatively complete methods, although we 
have never seen a practical ring-type erase head achieve this noise 
lev&l. The same remarks apply to such noise as may be generated by 
the bias field in the recording head, with one exception discussed 
below. We have found that any measurement of the erased noise 
level of a tape was in reality more nearly an evaluation of the erase 
and record heads and their electrical supplies. 




An erase head or record head which has a small permanent mag- 
netization will cause the tape to be noisy, and this- same effect can 
be caused by a small direct-current component or by an unsymmetrical 
wave form in erase or bias currents. If there are direct-current com- 
ponents in the erase field or if, in the extreme case, the tape is erased 
by a permanent magnet, the noise is increased; any irregularities in 
the tape cause irregular magnetization and the head, which is sensi- 
tive to the time rate of change of flux through it, picks up these ir- 
regularities. For this reason we have used a direct-current erase 
extensively in evaluating tapes since any shortcomings of the tape, 







I I 







X>2 05 .1 2. .512 5 



Fig. 6 Effect of tape irregularities of long wavelength in producing 
sidebands near a recorded signal. 

whether they be imperfect dispersion of the oxide in the binder or 
roughness of the front or back surface of the magnetic layer, are 
detected by this method. Furthermore, analysis of the noise into a 
frequency spectrum frequently indicates the source of noise and 
suggests the cure for it. The noise in direct-current wipe thus is a 
valuable tool in research but also is a measure, in some degree, of 
the noise which will result in practice from inevitable slight imper- 
fections in the recording system. It also has value as an easily re- 
produced condition which may be made the basis of a reference level 
in comparisons between various laboratories using different recording 


The direct-current noise is also of significance in indicating the 
"modulation" noise to be expected. This is the noise which occurs 
in the presence of a signal and which is absent in an erased tape. 
Chapin 5 first showed that the modulation noise near the frequency 
of a recorded note was composed of sidebands formed by the major 
components of direct-current noise. Fig. 6 shows an example of two 
tapes identical except for their direct-current noise. These tapes 
have -equal signal levels at the same harmonic distortion on our sys- 
tem. Fig. 6 shows the spectra of their direct-current noises, taken 
with a Hewlett Packard wave analyzer with 30-cycle half bandwidth, 
following an unequalized amplifier. The tape with a poor oxide 
dispersion has excessive low-frequency direct-current noise. It also 
shows the spectra resulting from recording a 400-cycle signal. It 
may be seen that the tape with low direct-current noise produces a 
purer tone and less background noise. The bandwidth plus certain 
wow in our equipment make the data somewhat inaccurate, particu- 
larly for the better tape, but the difference between the tapes is clear. 

This circumstance of noise resulting in the form of sidebands from 
the recording of a signal is the source of a possible noise caused by 
bias current, noted as an exception in the first paragraph on noise. 
While an ideal erasing head will subject the tape to gradually de- 
creasing fields and thus demagnetize it, a recording head must be 
built to have the transition from gap field to no field as abrupt as 
possible. Thus the bias frequency is normally recorded to some 
degree upon the tape, and irregularities in the tape will modulate the 
bias signal. Irregularities of wavelengths near that of the recorded 
bias signal may produce sidebands in the audio range. This effect 
in most recorders is less serious than effects of direct-current com- 
ponents or poor wave form, but is a form of limiting noise which can 
scarcely be avoided. It may, however, be reduced to insignificance 
by using such a high bias frequency that an insignificant level is re- 
corded. The frequency required will vary with tape, head, and 
speed, being proportional to speed. 

There are other forms of noise which are difficult to measure 
qualitatively but annoying to listen to, such as clicks and pops and 
such intermittent pulses which have been present in many different 
kinds of magnetic tape and which, likewise, are due to faults in con- 
struction of tape. They are not a necessary part of magnetic record- 
ing and may be eliminated by proper coating of a good dispersion 
upon an adequate backing. 


A final, rather special, form of noise is the signal irregularity caused 
by nonuniformity of the coating. This may be tested by recording a 
uniform signal and playing back through a suitable graphic recorder. 
In the past year, the uniformity of tapes, both within a single roll and 
from various production runs, has been much improved. Uniformity 
within 0.3 decibel in a roll and =*=0.6 decibel among many dif- 
ferent rolls may now be expected. 


The authors wish to thank the management of the Minnesota 
Mining and Manufacturing Company for permission to publish the 
data contained in this paper. 


(1) L. C. Holmes and D. L. Clark, "Supersonic bias for magnetic recording," 
Electronics, vol. 18, p. 126; July, 1945. 

(2) D. L. Clark and L. L. Merrill, "Field measurements on magnetic recording 
heads," Proc. I.R.E., vol. 35, pp. 1575-1580; December, 1947. 

(3) W. W. Wetzel, "Review of the present status of magnetic-recording theory," 
Audio Eng., vol. 32, p. 28; January, 1948. 

(4) L. C. Holmes, "Some factors influencing the choice of magnetic medium," 
/. Acous. Soc. Amer., vol. 19, pp. 395-403; May, 1947. 

(5) D. M. Chapin, "Measurement and calculation of under-signal noise in 
magnetic recording," Program 33rd Meeting Acoustical Society of America, May, 


MR. LEWIN: We made some tests at the Signal Corps using a Brush and regular 
tape, I think the type called No. 110, and we have experienced considerable 
trouble with what we call an echo. You referred to it as a print-through effect, 
an echo effect, from one layer to another, and so far, we have not been able to 
find what causes it. It is not always consistent. Sometimes we do not get it. 
Other times we get it at a fairly low recording level. 

MR. R. HERR: There will always be a certain transfer effect from one layer to 
another, but the fact that it does not occur all the time indicates that it is not 
necessarily associated with magnetic recording. It does vary with level, of course. 
The fact that it occurs only sometimes indicates that the roll on which it occurs 
has been handled in some way differently from the roll in which it does not occur. 
At least two factors operate to increase that effect. One of those is the presence 
of an alternating magnetic field. If you have an alternating field in the vicinity 
of a recorded tape, you will increase the transfer, the echo effect, enormously. 
Such fields are not likely to be encountered I am sure you have not deliberately 
placed the recording in a strong field, but even a weak field can have an appreci- 
able effect. Another factor is heat. A roll stored at high temperature will 
show the effect to a larger degree than a roll stored at low temperature. That is a 


sort of idealization process of an unrecorded layer in a field of recorded layers, 
and it is like a piece of iron in a weak field being hit with a hammer it will ac- 
quire more magnetization than without the mechanical strain. 

MR. LEWIN: We did try to expose a roll, before recording, to a strong field, and 
we used the voltage regulators which ordinarily gives quite a strong field, and 
did not notice any effect at all. 

MR. HERR: Before recording, you should not expect any effect from that, but 
once it is recorded, you have a modulated layer adjacent to another layer in the 
field of the recorded layer, a very weak field, then you apply an alternating field 
to this entire region, and the previously unrecorded layer is strongly magnetized, 
even by the weak field of the recorded layer. What you have done before you 
record it has no bearing on this. 

QUESTION: Sometimes you get it not from one roll to another, but in one part 
of a roll, and then you will not hear it again for the rest of the roll. 

MR. HERR: That I cannot explain. 

QUESTION: We have gotten it fairly consistently, with two different recorders 
and two different recording heads. 

MR. HERR: You say it is not related to the level? If it were related to the level, 
then there would be a fairly obvious explanation because the characteristic for 
such a field, in the absence of bias, is highly nonlinear, but except for that, I can- 
not explain it; the tape should be the same unless it was wound differently or 
treated differently in one place or another. 

QUESTION: An interesting effect we noticed was that if we did not rewind the 
reel after recording it, the print-through appears as an echo, as it followed the 
modulation. If we did rewind it, then it appeared as an anticipation. 

MR. HERR: I have noticed that effect myself. 

Wide-Track Optics for 
Variable- Area Recorders* 



Summary Wide variable-area sound tracks that meet Academy Re- 
search Council dimensional specification 1 RC-5001 have the advantage that 
they yield an improvement in ratio of signal to noise. The optical system 
for the new RCA Type PR-31 de luxe recording machine is equipped to 
make sound tracks to this standard. RCA recorders of earlier types can 
also be converted to record tracks to this standard. 

The optics that provide this feature are confined to the slit and objective 
lens assembly and involve no extensive changes in the optical system. 
They consist primarily of an objective lens of new design that provides addi- 
tional magnification in the direction of the slit length to meet the RC-5001 
standard. When systems of earlier types are converted, a new slit and ob- 
jective-lens barrel are installed, but no other optical changes are required. 
The converted system will produce all types of area tracks that could be 
made before conversion, and may be equipped for phototube monitoring. 

THE SIGNAL-TO-NOISE ratio of a sound track is improved by increas- 
ing the width of the sound track. If the noise originates pri- 
marily from a grainy structure in either the dense or clear portions of 
the sound track, this structure being characterized by a very large 
number of very small particles that are randomly distributed, it can be 
shown that the amplitude of reproduced noise will vary as the square 
root of the effective width of the sound track. At the same time the 
amplitude of the reproduced signal will vary directly as the effective 
width of the track. A net gain in signal-to-noise ratio of 3 decibels 
will therefore result from doubling the effective width of the sound 
track. The noise that is due to fog or a very dirty condition in the 
clear portions of a variable-area print will approximately respond in 
this manner to an increase in effective track width. Also the noise or 
hiss that accompanies the reproduction of a variable-density print 
responds in the same way. 

Another type of noise is found in sound tracks. It is of a different 
nature and is objectionable primarily when signal levels are very low 
or when no signal is present. Under this condition the transparent 
areas of the printed track are reduced to a minimum, or the density is 

* Presented May 20, 1948, at the SMPE Convention in Santa Monica. 


90 SACHTLEBEN January 

raised to a maximum, by the action of the noise-reduction system. 
Noise caused by film grain or dirt is therefore very low, but occasional 
minute pinholes, or small transparent areas in the print, due to dirt or 
other imperfections hi the negative, give rise to discrete sounds vari- 
ously known as "ticks," "pops," or "crackles." These sounds are 
individually recognizable because they occur relatively infrequently, 
as compared to the sounds that occur frequently enough to produce 
the continuous effect of hiss. For this reason they constitute an ob- 
jectionable residue of noise under conditions that ideally should be 

The effect of widening the sound track is to reduce the relative 
amplitude of these discrete "pops" even more than hiss. Doubling 
the track width doubles the number of discrete sounds per second but 







Fig. 1 Dimensions of wide-track variable-area record- 
ing slit image and wide-track reproducing slit image ac- 
cording to Academy Research Council Specification 1 

has no effect on their absolute amplitude as long as they occur sepa- 
rately. The amplitude of the desired signal is, however, doubled. 
The ratio of signal-to-discrete-noise level is therefore increased 6 deci- 
bels. The fact that twice as many "pops" occur, each of a level that 
is relatively 6 decibels lower than before, is of no consequence as long 
as their probability of simultaneous occurence is still substantially 
zero. Sound tracks that are wide enough to double the amplitude of 
the useful signal effectively are therefore especially effective in sup- 
pressing residual surface noise of this nature at very low recorded 

Fig. 1 illustrates the present 35-mm industry standard dimensions of 
the wide variable-area sound track. The negative track is recorded 
in two 0.076-inch-wide halves, with a separation of 0.024 inch between 
them to facilitate isolation of the halves in push-pull reproduction. 
The print is scanned to a total width of 0.184 inch, in two 0.084-inch 
illuminated segments separated by a 0.016-inch septum. As compared 


with the 35-mm standard narrow track, the recorded signal will be 
greater in the ratio 0.152/0.076, the reproduced hiss from the trans- 
parent part of the track will be greater in the ratio V 0.152/0.076, and 
the level of the "pops" will be unchanged. The hiss from the dense 
parts of the print at densities in the range 1.3 to 1.5 is so low as to be 
negligible for tracks of both widths. 

Special optics have been designed that make any RCA variable-area 
studio recording optical system, of the PR-23 and later types, con- 
vertible for recording the wide sound tracks. The optical system for 

Fig. 2 Wide-track conversion assembly ready to install 
in narrow-track optical system. 

the new PR-31 de luxe recorder is also available equipped with the 
special optics for wide-track recording. These optics are contained 
within the barrel or assembly that houses the slit and final objective 
of the system. Fig. 2 shows the appearance of a conversion assembly 
ready to install. 

In addition to the optical changes, the distance from the slit to the 
film must be increased about 1/8 inch by relocating the optical sys- 
tem, and the system must also be moved laterally 0.050 inch to per- 
mit one half of the wide track to be located in the standard position, 
with its center 0.243 inch from the edge of the film. This permits one 
half of the track to be reproduced in any standard soundhead. 

92 SACHTLEBEN January 

Inasmuch as the wide-track technique is intended primarily to im- 
prove the quality of original sound tracks which will later be re- 
recorded to standard release negatives, it is advisable that wide tracks 
be of the push-pull type. Variable-area push-pull sound tracks in 
common use are of three types, known as Class A, Class B, and Class 
AB. Each has its own peculiar characteristics and advantages from 
the operational standpoint, but all have quality advantages over any 
of the tracks of the nonpush-pull type. 

The special wide-track optics were first designed with the conver- 
sion of existing optical systems in view. To this end details were 
worked out in such way that any of the push-pull apertures that had 
been designed for the PR-23, or later studio recording optical sytems, 





"^ I r .OOO25* 



Fig. 3 Schematic of the modulating branch of the optical 
system showing wide-track optics. 

could be used without change in the production of wide tracks. As 
a result the wide tracks are fully modulated at a galvanometer ampli- 
tude that is about 1 decibel below the corresponding amplitude for 
narrow-track push-pull recording. In the case of the Type PR-31 
recorder optical system, apertures have been especially designed for 
wide-track push-pull work, and this optical system fully modulates 
all types of sound tracks at a uniform galvanometer amplitude. 

Fig. 3 shows two schematic views of the modulating branch of a 
variable-area recording optical system, that extends from the galvanom- 
eter mirror to the film, and illustrates the nature of the wide-track 
optics. The optics in all other parts of the optical system are the 
same for either wide-track or narrow-track work. In the figure, A 
is the galvanometer mirror, B is the slit condenser, and C, the slit. 
H is the film. Parts Z>, E, F, and G constitute the special wide-track 


optics which are combined to form a well-corrected anamorphote 

In the transverse plane (plan view of the figure), lenses D and E 
constitute an air-spaced cylindrical lens which co-operates with the 
cemented spherical doublet F to image the slit upon the film. Cylin- 
drical lens G has no power in the transverse plane. In the longitu- 
dinal plane (elevation view of the figure), cemented spherical doublet 
F co-operates with cemented cylindrical doublet G to image the slit 
upon the film. Cylindrical lenses D and E have no power in the 
longitudinal plane. The reduced image of the slit at the film is 
formed at different magnifications in the two planes, so that the ratio 
of its length to its width is much greater than the corresponding ratio 
for the slit itself. This ratio is about 2.44 times as great for the image 
as for the slit, and it is by this inequality of the two ratios that the slit 
image is made long enough to record a track that is 0.176 inch wide. 
The anamorphote objective is corrected for all the usual aberrations in 
both planes. The air-spaced cylinders D and E are designed to com- 
pensate the curvature introduced by lens F } and the result is that the 
wide-track optical system exhibits no more image curvature than a 
narrow-track system. 

Anamorphote-lens combinations require critical adjustment of the 
azimuth of the cylindrical components in order to perform properly. 
Special fixtures and procedures are employed to align the lenses by 
optical test, and the mountings are especially designed to hold the 
lenses securely in the position so determined, after they are assembled 
into the optical system. 

Since the increase in track width has been obtained without any 
increase in the area of the slit, this obviously results in exposure of the 
wide-track negative falling to about 40 per cent of the exposure for a 
narrow-track negative. Compensation for this exposure loss is made 
by employing fine-grain stocks similar to Eastman 1372 and exposing 
them with white light. Adequate negative densities are obtainable, 
and the quality of the sound-track image is substantially equivalent 
to that obtained with ultraviolet exposure and emulsions similiar to 
Eastman 1357. 

Fig. 4 illustrates the action of the modulator and noise-reduction 
shutter, in the Type PR-31 recorder optical system equipped for 
recording Class A push-pull wide track. A drawing that illustrates 
the character of the negative track produced is also shown. The slit 
S and aperture images A and B, which are in the plane of the slit, are 




shown as they would be seen to appear from the position of the galva- 
nometer mirror. The dotted line V is the outline of the noise-reduc- 
tion shutter vane located just beyond the plane of the slit, but not 
visible. An opaque portion or septum SP occupies the center of the 
slit. The slit is thus divided into two active segments, one being 
illuminated by aperture image A and the other by B. The cutting 
edges of these images cross their respective segments of the slit at 
points midway between their centers and ends, when no modulation is 

Fig. 4 Action of the Class A push-pull 
aperture and slit in the Type PR-31 re- 
corder wide-track optical system. 

present, and the central half of each segment of the slit thus is fully 
illuminated. Two edges of the opening in the shutter vane also 
cross each active segment of the slit and block all of the light entering 
the slit from A and B with the exception of four small equal portions 
C. In the accompanying drawing of the negative sound track, the 
four strips or bias lines L are produced by the portions C of the slit 
that are illuminated by A and B, but are not covered by the shutter 

As modulation currents are applied to move images A and B up and 
down together, a motor controlled by the modulation currents causes 
the vane V to move downward to uncover sufficient additional por- 
tions of the active segments of the slit to accommodate the excursions 




of A and B. When modulation currents cease, the condition shown in 
the illustration is resumed. An auxiliary image M moves with images 
A and B. This image lies in a window W in the plane of the slit, and 
its top portion is limited by the tab T on the shutter vane. When 
the image M is reimaged on the monitor screen, it provided a means 
for observing the displacements of images A and J5, and also of the 
shutter vane V. 

The Class A push-pull sound track produced by this arrangement is 
in two parts, J and K. Each of these is a normal duplex track with 
noise reduction, and can be played by itself as a complete record. 
The two tracks J and K are in 180-degree phase relation to each other. 
The push-pull track has the advantages that the noise-reduction 
envelopes of the portions / and K are in phase and cancel in push-pull 
reproduction, and that even-harmonic distortion of photographic 
origin also cancels. The use of push-pull is thus essential when mak- 

Fig. 5 Typical wide yariable-area sound track Class A 
push-pull print (recorded on converted optical system). 

ing original negatives of the highest possible quality. Portion H 
illustrates overshooting. 

Fig. 5 shows a section of a typical Class A wide-track push-pull 
print of recorded speech. This was made with a narrow-track optical 
system that has been converted to record wide sound tracks. 

Studio operational experience with optical systems con verted to wide 
track indicates their frequency-response performance is comparable 
with narrow-track systems. No special compensation was required 
by the conversions. Over-all distortion measured from playback 
negatives, including distortion of signal generator and recording and 
reproducing channels, is as follows : 

Level Referred to 
100 Per Cent Modulation 

2 decibels 
4 decibels 
8 decibels 


2 . 4 per cent 
2 . per cent 
1 . 8 per cent 


The reduction in relative level of the "ticks" and "pops," which are 
the most serious noise at very low levels or in silent parts of variable- 
area records, has been very satisfactory. 

Exposure characteristics follow: 

Lamp Current 7.2 amperes in 7.8-ampere lamp 
Negative Density 2.7 
Print Density 1.3 to 1.5 
Negative Stock Eastman 1372 
Processing Commerical laboratory 


Acknowledgment is due the help of many people who participated 
in developing the wide-track optics. Credit is particularly due Mr. 
G. L. Dimmick, Mr. J. L. Pettus, Mrs. Mary Smuck, who did the 
computing, and Mr. E. S. Leslie, who responsibly accepted and solved 
many of the practical problems of assembling the first units that were 
built. The requirement for cylindrical lenses of adequate quality for 
this design was met by the Herron Optical Company of Los Angeles. 


(1) Academy Research Council Specification RC-5001 is now incorporated in 
American Standards Z22.69-1948 and Z22.70-1948 (Universal Decimal Classifica- 
tion *UDC 778.534.4), /. Soc. Mot. Pict. Eng.,'vo\. 51, pp. 547-548; November, 

Trend Control in 
Variable-Area Processing* 



Summary This paper compares two alternate methods of obtaining a 
signal for cross-modulation testing. It also describes a meter which will 
give from a single print density the information formerly given only by a 
series of prints of different densities. This meter will be of help in pre- 
dicting trends in processing variations due to changes in developing. 


As POINTED out by Baker and Robinson, 1 a cross-modulation test 
affords an extremely accurate means of determining correct 
negative and print densities for given conditions of laboratory proc- 
essing for a variable-area sound track. This type of test is now in 
universal use. The modulated test tone consists usually of a 9000- 
cycle carrier, amplitude-modulated by 400 cycles The peak ampli- 
tude of this modulated wave is adjusted to about 90 per cent of a 
fully modulated track. A 1000-cycle tone is recorded as part of the 
test to be used as a reference level, its amplitude being the same as the 
peak amplitude of the modulated wave. The exposure on the sound 
negative is held to such a value that, when the negative is printed onto 
a positive film and developed normally, minimum cross-modulation 
products will be present. In order to determine the proper printing 
exposure (the developing time is held to a certain value fixed by the 
picture requirements) several prints of different densities have to be 
made. The cross-modulation products are measured through a 400- 
cycle band-pass filter and are plotted, in reference to the 1000-cycle 
tone, against print densities. From the curve one selects the proper 
printing point, which is optimum cancellation. This optimum, which 
usually measures below the noise level of the system, can never be 
reached unless the oscillator generating the modulated wave has an 
output which is free of the 400-cycle, the modulating frequency. Few 
laboratories have such an oscillator available, and fewer are equipped 
to build one. 

* Presented May 18, 1948, at the SMPE Convention in Santa Monica. 





From the mathematics of an amplitude-modulated carrier wave, 
generated in a push-pull modulator, one can see that the alternating- 
current output wave consists of a carrier and two sidebands. Assumed 
in this statement is that the cancellation of the modulation frequency 
is complete, i.e, that the circuit is balanced and that either the galva- 
nometer and/or the film will work as a low-pass filter capable of 
suppressing other components generated in the process of modulation. 
Fig. 1 shows a microphotograph of a cross-modulation test made with 
a Radio Corporation of America cross-modulation oscillator. The 
film used was Eastman 1372 exposed with white light and developed 
in high contrast sound-track negative developer. 

Fig. 1 


The cross-modulation test in itself does not require a signal which 
must contain a carrier and two sidebands. 2 The test can be made with 
identical results by using a single-sideband modulated carrier. In 
this case, instead of modulating a high frequency by 400 cycles, it is 
easier to mix two high-frequency signals in a linear network (resist- 
ances) which have a different frequency equal to the modulating 
frequency. Two signals of equal amplitude will give a 100 per cent 
modulated wave, i.e., they will add to zero when opposite in phase and 
to twice the peak amplitude when in phase. The single-sideband 
frequency can be either the carrier frequency plus or minus the modu- 
lating frequency, i.e., in the case of 9000; 8600 or 9400 cycles. As 
present-day standards call for only 80 per cent modulation, the side- 
band should have only 80 per cent amplitude in reference to the car- 
rier. Fig. 2 shows a microphotograph of a single-sideband modulated 
carrier signal. Again Eastman 1372 film was exposed with white 
light and developed in high-contrast sound-track negative developer. 




With phase-shift oscillators, such as manufactured by General 
Radio or Hewlitt-Packard, one will have no trouble in keeping the 
difference in frequency sufficiently constant. As the signal is applied 
at a very low level it is advisable to build the two oscillators out 
through high-loss pads, before combining the two signals to eliminate 
interaction. If the harmonic contents of both oscillators are high it is 
advisable to feed one of the signals through a low-pass filter. Usually 
a single-section, constant-K filter, 3 with a cutoff frequency equal to the 
oscillator frequency is sufficient. Under these conditions, the writer 
has measured more than a 70-decibel (the limit of the test equipment) 
differential between the wanted and unwanted signal. 

Fig. 2 


Under certain conditions, as in print production, it is impossible to 
make a series of print densities. From the electrical measurement, 
one can only determine if the print falls within the limits given, 
usually 30 decibels relative to 1000 cycles on either side of the mini- 
mum. If the cross-modulation products, for some unforeseen reason, 
should exceed the limits, visual means are at present the only recourse 
for determining if the print is too dark or light. Trends within the 
limits, which, if known, may prevent future excessive cross modulation 
can hardly be determined by the visual method. 

In Fig. 3 there is shown a meter built for a different purpose that 
may be of help in determining trends in processing a variable-area 
sound track. The signal is picked up and amplified by usual means. 
The input to the meter is adjusted by PI so that the carrier amplitude 
measured by MI is +10.0 decibels relative to 1 milliwatt across 600 
ohms (0 dbm*) . Part of this signal is fed through a unity gain ampli- 
fier into a ring modulator. 

* Decibels with respect to 0.001 watt. 




Part of it through a 15-decibel, 600/600-ohm isolation pad, a single- 
section 400-cycle (f/fm = 1.2) constant-K band-pass filter 4 to an 
amplifier having 42 decibels gain. In setting up the equipment, the 
gain from the input terminals of the isolation pad P* to the output of 
the amplifier should be adjusted to 25 decibels. That is, so M 3 can be 


M 3 


. 246 microfarad 
3 . 58 microfarads 
0.645 henry 
43 . 8 millihenries 
+ 10 dbm* at zero of scale 
500-0-500 microamperes, direct cur- 

+5 dbm* at zero of scale 
600/600-ohm 15-decibel pad 
Ft 100-ohm slide wire 

P 3 20-decibel .2-decibel step meter pad 

P 4 600/600 ohms 15-decibel pad 

RiRz 600 ohms 

C U ring modulator 

Tt 500/500 repeat coil 

* Decibels with respect to 0.001 watt. 
Fig. 3 

calibrated to measure the cross-modulation products; i.e., the 400- 
cycle tone contained in the original signal, directly without applying a 
correction factor. M s has a sensitivity of +5.0 dbm at zero on its 
scale and is calibrated to 10.0 relative to its own zero. An auxiliary 
potentiometer in series with M 3 decreases the sensitivity of the meter 




by 20 decibels in 2-decibel steps. Thus M" 3 will read, if MI is adjusted 
to "0" at the 1000-cycle reference signal, cross-modulation products 
of either a single- or a double-sideband 400-cycle modulated signal 
from -10.0 to -40.0 directly. 

The 400-cycle tone is then fed across the other two terminals of the 
ring modulator. To balance the ring modulator after construction of 
the meter, a 9000-cycle tone at a level of +10 dbm is sent into the 
instrument and PI (a slide wire) is adjusted until M 2 , the phase meter, 
does not give an indication ; i.e., the ring modulator is balanced. This 
adjustment does not have to be reset unless the ring modulator be- 
comes defective and a new one has to be installed. 






180 360 

A modulated signal, which does not contain cross-modulation prod- 
ucts, will behave similar to a single tone ; that is, the ring modulator 
stays balanced and therefore M 2 reads zero. If the modulated signal, 
hereafter called EI (Fig. 3) contains cross-modulation products, the 
400-cycle tone will pass through the 400-cycle band-pass filter unim- 
peded, is amplified and will appear as E% across meter MZ, and is subse- 
quently fed into the ring modulator. E 2 will unbalance the ring modu- 
lator one way or the other, depending on the phase of the voltage E 2 
with respect to voltage EI. This unbalance will result in a positive or 
negative indication of meter Mz, the magnitude of which will depend 
on the amplitude of E z as EI is fixed. As the rectifiers or diodes are 



connected in full wave, the detector operates over the full cycle. The 
output is nearly linear for variation of E 2 . 

Fig. 4 shows the resulting wave forms of a cross-modulation test 
when under- , correctly, or overexposed. It also makes it evident that 
E 2 is always either in phase or 180 degrees out of phase with EI. If E l 
is held at a constant amplitude and E 2 is either in phase (Fig. 4-C) or 
180 degrees out of phase (Fig. 4-A) with EI, I M z (Fig. 3) can be cali- 
brated for a given laboratory condition directly in under- or overex- 
posure in terms of "lights" or density. After calibration, the phase 
output of the reproducer or amplifier is of no account as long as the 


see text 

+8.0 dbm* at zero of scale 4000 


46.5 decibels 600/600 ohms 
200 ohms 
1000 ohms 

* Decibels with respect to 0.001 watt. * 

Fig. 5 

phase relationship of EI to E* inside the meter stays the same. The 
phase sensitivity of the meter can be altered to suit particular condi- 
tions by inserting an attenuator between M 3 and T 2 . 

Fig. 5 shows the circuit diagram and values for the cross-modulation 
single-sideband generator. 


(1) J. O. Baker and D. H. Robinson, "Modulated high-frequency recording 
as a means of determining conditions for optimal processing," J. Soc. Mot. Pict. 
Eng.-vol. 30, pp. 3-18; January, 1938. 

(2) A. Narath, "Entstehung und Beseitigung des Donnereffektes bei Zack- 
centonfilmen," Zeit.fiir Tech. Phys., vol. 5, pp. 121-130; May, 1937. 

(3) T. E. Shea, "Transmission Networks and Wave Filters," D. Van Nostrand, 
New York, N. Y., October, 1929, p. 225. 

(4) See p. 230 of reference 3. 


Semiannual Convention 

rriHE 64TH SEMIANNUAL CONVENTION of the Society of Motion 

L Picture Engineers was held at the Hotel Statler in Washington, 
D. C., October 25-29, 1948. There were 400 members and guests 
who registered for the seven technical sessions and the two symposia 
on high-speed photography. The Banquet was attended by 265, 
and there were 209 at the Luncheon. 

Eric Johnston, president of the Motion Picture Association was 
the Guest Speaker at the Luncheon. John Russell Young, presi- 
dent of the Board of Commissioners, Washington, District of Colum- 
bia, also spoke. 

At the Banquet, Mr. Ryder presented the Progress Medal Award 
to Peter Mole, and the Samuel L. Warner Award to Nathan Levin- 
son. Sixteen Active Members were elevated to the grade of Fellow. 
Mr. Ryder also introduced Mr. Sponable, the new president of the 

On Thursday evening, a special session was held at the Naval 
Photographic Center. This included a tour of the Center, and was 
followed by a technical session held there. 

Mrs. Nathan D. Golden, hostess for the Ladies' Committee, 
prepared an interesting program of sightseeing for the women guests. 


ON OCTOBER 28, 1948, Mr. Peter Mole was presented with the 
1948 Progress Medal Award, given for outstanding achieve- 
ment in motion picture technology. Mr. Mole was chairman of the 
Pacific Coast Section of the Society shortly after it was formed in 
Hollywood. He has also been a member of the Board of Governors 
and is Executive Vice-President of the Society for 1949-1950. 

Born in Termini, Sicily, Italy, he was brought to the United 
States when he was six years old. After being educated in various 
technical schools he joined the engineering staff of the General 
Electric Company at Schenectady, New York, where he was active 
in the development of the General Electric searchlight and a high- 
intensity rotating carbon-arc theater projection lamp. 

In 1923 he left the General Electric Company and moved to 
California where he became interested in motion picture studio 
lighting, first with the Metro-Goldwin-Mayer Studios in the elec- 
trical department. After receiving his groundwork training in 
actual production, he went to work for a motion picture studio 
lighting equipment manufacturer. 

With his technical background plus experience in the studios he 
was ideally suited to enter the field of the manufacture of specialized 
equipment for an industry that was growing so fast its requirements 
changed almost from month to month. 

It was not long before he joined forces with Elmer C. Richardson, 
another design engineer, and Fielding C. Coates, a studio chief 
electrician, and formed the Mole-Richardson Company. 

Mr. Mole and other members of his organization have been the 
authors of numerous papers on motion picture studio lighting which 
have been published in the JOURNAL, as well as serving on many 


Society committees. His technical contributions and those of his 
organization are outstanding. 

Mr. Mole's success in his chosen field is not due entirely to his 
ability to organize and operate an engineering and manufacturing 
organization to meet the needs of a unique industry. He has an un- 
usual insight into the intangibles created by the art form in motion 
picture production. He knows that engineering perfection must not 
transcend utilization in an industry where dramatic effect is the end 
result; yet he has been able to design and produce highly specialized 
lighting tools which satisfy both the artist and the engineer. It is 
for these reasons that, 

"It is the unanimous recommendation of the Progress Award Committee that 
the Progress Medal Award of the Society of Motion Picture Engineers be 
awarded this year to Mr. Peter Mole for his many and continuing contribu- 
tions to Motion Picture Lighting technique and equipment. 

"Mr. Mole and his organization have pioneered in the development of 
lighting equipment during each of the successive stages in the advance of 
motion picture lighting practice for more than twenty years. The wide use of 
their products manifests the success of their efforts and achievement. Future 
lighting developments which are currently being studied undoubtedly will re- 
flect the benefits of Mr. Mole's experience as new and improved lighting be- 
comes available for studio use. 

"It is fitting to note that Mr. Mole's organization has been the recipient of 
four Certificate Awards from the Academy of Motion Picture Arts and 
Sciences, together with recognition by the United Nations Conference on In- 
ternational Organization." 


WARNER BROTHERS established the Samuel L. Warner 
JL Memorial Award, to be given each year by the Society of 
Motion Picture Engineers at its Fall Convention to an engineer 
selected by the Society, who has done the most outstanding work in 
the field of sound motion picture engineering. The 1948 Award was 
presented to Nathan Levinson who has had a long and successful 
career in radio communications as well as in sound motion pictures. 
He started his radio work as an engineer for Marconi, prior to the 
First World War, and served in the United States Army Signal 
Corps during that war, rising to the rank of major. After the war he 
joined the Bell System as a commercial engineer in the radio broad- 
cast field. 

Shortly after this, when sound was proposed as an adjunct to 
motion pictures, various attempts were made to interest the studios 
in sound for motion pictures, but only Samuel L. Warner was con- 
vinced of its commercial possibilities. The history of the growth of 
sound in motion pictures is a matter of record and need not be re- 
peated here. 

During World War II the Navy asked the Warner Brothers to 
take over the manufacture of a special combat camera, after others 
had failed. Mr. Levinson undertook this responsibility in addition 
to directing Warner Brothers' Sound Department and successfully 
produced and delivered these cameras. 

He pioneered such ideas as the intercutting of variable-area and 
variable-density sound tracks for increased volume range, the com- 
mercial use of control track for increased volume range, and one of 
the first camera blimps, permitting the cameraman to come out of 
the soundproof camera booth imposed by the advent of sound 

The use of 16-mm motion pictures with high-speed development, 
while not an original idea with Mr. Levinson, was, under his 
guidance, commercialized for recording race-track events. Cur- 
rently, Mr. Levinson is playing an important role in the develop- 
ment of television for theater use and as a tool for the production of 
sound motion pictures. He is also active in the commercial de- 
velopment of a new color system for motion pictures. 

The industry is proud of Mr. Levinson and owes to him a debt of 
gratitude for his many technical contributions to the advancement of 
the art. 



The Journal Award Committee recommended that the Award 
for the year 1948 be given to Messrs. J. S. Chandler, D. F. 
Lyman, and L. R. Martin for their paper, "Proposals for 16-mm and 
8-Mm Sprocket Standards/' published in the June, 1947, issue of the 
JOURNAL! This deals with the design of that fundamental motion 
picture mechanism, the sprocket wheel, and the variables that affect 
the interaction of the film and the sprocket. 

They have also proposed standard sprocket-design formulas which 
allow the engineer to accommodate the variables that apply to his 
particular problem, being adaptable to any application regardless 
of the size and function of the sprocket or the path and shrinkage of 
the film. If changes are made in the physical properties of the film 
or if research discovers conditions of improved operation, the 
formulas can be adjusted. 

Dr. J. S. Chandler was born in Nebraska and attended the 
Georgia School of Technology, obtaining a B.S. degree in Mechanical 
Engineering in 1934 and an M.S. degree in 1936. He became a Re- 
search Fellow at the Pennsylvania State College the following year 
and received his Ph.D degree in Mechanical Engineering in 1938. 
Since that time he has been employed in the Sound Recording Sec- 
tion of the Physics Department in the Kodak Research Laboratories, 
where his work has included the studies of the mechanical compo- 
nents of cameras, printers, and other equipment in connection with 
sound film. He is the author of several previous papers, including 
one published in the JOURNAL of the Society of Motion Picture Engi- 
neers dealing with design considerations of film sprockets. 

Mr. Donald Franklin Lyman was born in Winsted, Connecticut. 
He was graduated from Massachusetts Institute of Technology in 
1921 with the B.S. degree in Mechanical Engineering. After em- 
ployment with the Western Electric Company for three years, he was 
engaged in 1924 as special engineer in the Development Depart- 
ment at the Camera Works division of the Eastman Kodak 

He has worked on fire-control apparatus for airplanes, has done a 
considerable amount of work for the American Standards Associa- 
tion's Committee on Motion Pictures, and at present, his work in- 
volves the development and engineering inspection of amateur 
motion picture apparatus. 

Mr. Lawrence Randall Martin is assistant to the manager of the 
Camera Works division of the Eastman Kodak Company and has 
been associated with the company since July 6, 1931. 

Mr. Martin was first employed as a draftsman; two years later 
he became a product designer, and subsequently chief engineer on 
motion picture cameras. 

In January, 1940, he was named production technician, represent- 
ing the company in contracts with the Sperry Gyroscope Corpora- 
tion, the Ford Motor Company, and the United States Army Ord- 
nance Division. He later joined the manager's office to act as 
liaison between the company and the Armed Services. 

At the present, Mr. Martin is responsible for co-ordination of 
efforts of all departments in new products programs. 

He was born in Punxsutawney, Pennsylvania, and received the 
Mechanical Engineering Degree from Cornell University in 1931. 

He has served on several committees of the Society and is par- 
ticularly interested in the industrial applications of motion pictures. 

Fellow Awards 1948 

AT THE BANQUET held on October 27, 1948, President Ryder 
presented sixteen Active Members of the Society with the 
Fellow Award. The names of the recipients are as follows: 


Radio Corporation of America National Carbon Company 


Ansco Corporation Eastman Kodak Company 


George W. Colburn Laboratories Bell Telephone Laboratories 


Western Electric Company Neumade Products Inc. 


Radio Corporation of America Western Electric Company 


Massachusetts Institute of Technology Eastman Kodak Company 


Allen B. DuMont Laboratories, Inc. Walt Disney Studios 


Ansco Corporation Eastman Kodak Company 

Section Meetings 


The October 22, 1948, meeting of the Central Section of the SMPE was held in 
the Auditorium of the Engineering Building in Chicago, in joint session with the 
Chicago Section of the Institute of Radio Engineers. 

Kenneth Jarvis of the IRE called the meeting to order and introduced R. T. 
Van Niman who, in turn, introduced the first speaker, Ernest F. Zatorsky, director 
of sound recording, The Jam Handy Organization, Detroit. 

His paper titled "Microphone Placement Techniques as Applied to Motion 
Picture Sound Recording" outlined problems involved in securing good quality 
sound pickup without having the microphone appear in the picture. He advo- 
cates use of one microphone directly above the camera line and in front of the 

The next paper was "Synchrolite for Television Film Projectors" by L. C. 
Downes, Television Engineering Section, General Electric Company, Syracuse. 
This paper described a pulse light source for a standard projector operating at 24 
frames per second with the shutter removed, the pulse rate being 30 per second to 
synchronize with tube scanning. The lamp is Krypton filled. The flash points 
are a tungston alloy and arc at about 70 to 80 amperes. Operation of the unit 
consumes 400 to 500 watts, and the life of the lamp is rated at 50 hours. The 
temperature at the film gate is very low, and the light delivered to the television 
pickup is 50 foot-candles. The pulse circuits were described in detail, and open 
discussion followed this presentation. 

The November 12, 1948, meeting of the Central Section was called to order by 
R. T. Van Niman in the rooms of the Western Society of Engineers. About 110 
members and guests were present. Short reports on the Washington convention 
and the election of national officers were given. 

"Carbon-Arc Projection," a technicolor film produced by National Carbon Com- 
pany was presented first. Preliminary comments were given by C. E. Heppberger 
of this Company. The film was projected with a 16-mm carbon- arc machine, 
and problems in the making of the film and techniques used were explained. 

"A Discussion of High-Quality Sound Reproduction" was given by John K. 
Hilliard, of the Altec Lansing Corporation. 

Mr. Hilliard outlined the requirements for high-quality sound reproduction 
from both the objective and subjective points of view, pointing out that an over- 
all flat system frequency response, the supposed ideal from the objective point of 
view, does not necessarily produce a satisfactory illusion of reality in the repro- 
duced sound. He discussed many of the factors which are believed to be respon- 
sible for this situation such as level differences between original and reproduction, 
various types of distortions introduced acoustically, electrically, mechanically, or 
photographically, and psychological conditioning of listeners. The considerations 
upon which the motion picture industry's optimum electrical-response curves for 
theater reproducing equipment are based were outlined, and Mr. Hilliard stated 
that similar curves for 16-mm reproducing equipment are being plotted. 


Section Meetings 

Acoustical-response measurements on typical good-quality theater equipment 
were presented, and dynamic power-level curves were shown which indicate that 
all components of sound systems must be capable of handling with low-distortion 
transients of extended frequency range and peak powers many times the so- 
cilled "normal" system ratings if the reproduced sound is to more than superfi- 
ciallyresemble the original. 

Mr. Milliard's formal presentation was followed by a discussion period of almost 
equal length participated in by many members of the Chicago audio group who 
attended the meeting. Spirited arguments developed as to the precise meaning of 
some of the terms used in discussing "high-quality" sound reproduction, and many 
dissenting opinions were expressed regarding such conclusions as have so far been 
drawn in the audio field. The one point of general agreement appeared to be that 
only a beginning has been made in reproducing sound, and that a great deal of fur- 
ther research and development is needed, particularly with respect to the psycho- 
logical aspects of the general problem. 

Pacific Coast 

An audience of approximately 150 members, admitted by membership card 
only, filled the Western Electric Review Room to witness a program consisting of a 
symposium of papers on "The Problem of Sound Reproduction on 16-Mm Koda- 
chrome," at the October 12, 1948, meeting of the Pacific Coast Section. R. G. 
Hufford of the Eastman Kodak Company presented a paper on the "Sensitometric 
Characteristics of the Kodachrome Sound Track," which was followed by a short 
sound demonstration film. Robert V. McKie of the RCA Victor Division gave 
a short talk on the establishment and maintenance of commercial processing 
tolerances for making variable-area Kodachrome sound track. J. G. Frayne of 
the Electrical Research Products Division of Western Electric outlines the 
methods of making variable-density sound track on Kodachrome and introduced 
a new system of "electrical printing" whereby the sound track is re-recorded onto 
each Kodachrome release print. Dr. Frayne gave a demonstration of the quality 
obtained by this latter method. 

The November 9, 1948, meeting of the West Coast Section was held at the re- 
cently completed broadcast studio of Radio Station KHJ and the Mutual-Don Lee 
network. Approximately 350 members and their wives attended this very inter- 
esting and informative evening's entertainment and tour through the studio. 

The studio management arranged to permit the audience to see a live- talent tele- 
vision broadcast as the opening phase of the evening's program. A portion of the 
audience saw the program reproduced on a large-screen television receiver in an 
adjacent auditorium in the studio. Following this television activity, W. 
Carruthers, chief engineer of Station KHJ, and F. L. Hopper of Western Elec- 
tric, discussed some of the outstanding architectural, acoustical, and electrical 
features of this studio, and Harry Lubcke, director of television at Station KHJ, 
discussed some of the interesting aspects of the television experiences of this 

The remainder of the evening was devoted to a conducted tour in small groups 
through the entire studio, with engineering experts available for questions from 
the audience concerning the engineering features of the equipment in the studio. 

Meetings of Other Societies 

Inter-Society Color 
Council Meeting 

The seventeenth annual meeting of 
the Inter-Society Color Council will 
be held on Wednesday, March 9, 1949, 
Hotel Statler, New York City. The 
meeting will consist of a Discussion 
Session at which committee chairmen 
will report on the following problems: 

2 Color Names (Revision of), 
Deane B. Judd 

6 Color Terms, Sidney M. New- 

7 Color Specifications, Walter C. 


12 Studies of Illuminating and 
Viewing Conditions in the 
Colorimetry of Reflecting Ma- 
terials, Deane B. Judd 
14 A Study of Transparent Stand- 
ards Using Single-Number 
Specification, Robert N. Os- 

A Business Session will conclude the 
afternoon meeting. Anyone interested 
is invited to attend. Hotel reservations 
should be made directly to the hotel at 
least ten days prior to the meeting, 
indicating that you are attending the 
ISCC meeting. 

Meetings of the Optical Society of 
America are scheduled for the same 
hotel, March 10-12. It is usual for one 
or more O.S.A. sessions to be devoted 
to color, and it is therefore suggested 
that all ISCC members who are in- 
terested plan to remain for these meet- 

PSA Convention 

The high light of the recent conven- 
tion of the Photographic Society of 
America, held in Cincinnati, Ohio, 
November 3-7, 1948, with more than 
700 photographers and technicians in 
attendance, was an all-day clinic of the 
Technical Division on "Photography in 

Among the outstanding technical 
papers presented were those by J. I. 
Crab tree, "Rapid Processing of Films 
and Papers"; Edward H. Loessel, 
"Making Duplicates of Color Trans- 
parencies"; H. G. Morse, "High-Speed 
Flash Photography in Black and White 
and Color"; H. A. Miller, "Direct 
Positive Transparencies by Chemical 
Reversal"; Harvey P. Rockwell, 
"Light Measurement in Photography"; 
and Allen Stimson, "Exposure and 
Light Measurement." 

The Motion Picture Division of PSA 
presented a number of interesting 
amateur and commercial films in six 
well-attended sessions. Among the 
papers dealing with motion pictures 
were: "Title Backgrounds by the 
Experts," by Dennis R. Anderson; "A 
Challenge to Your Talents," by Mrs. 
Warner Seely; "Electrical Remote- 
Control Unit for Movie Cameras," by 
Belgrave F. Gostin; "Home Movies in 
Agricultural Education," by George F. 
Johnson; "How to Make a Movie," by 
Charles H. Coles; "Making Movies of 
Football," by Harris B. Tuttle; ^Direct 
16-Mm Productions," by Larry Sher- 
wood; and "Photometric Calibration 
of Motion Picture Camera Lenses," by 
M. G. Townsley. 


Book Reviews 

Informational Film Year Book, 1948 

Published (1948) by the Albyn Press, 42 Frederick St., Edinburgh, 2, Scotland. 
200 pages. 21 figures. 5 3 / 4 X 8 3 A inches. Price, 12s. 6d. net. 

This second volume of the Informational Film Year Book follows the same pat- 
tern as the initial volume published last year. However, many additional data 
have been included making the book of even more value than before to producers 
and consumers of nontheatrical motion pictures. Eight feature articles appear on 
various phases of documentary films, filmstrips, and equipment. The remainder 
of the book consists of listings and directories covering films of the year, who's who 
in documentary, various organizations and societies the world over, film producers, 
special service firms, and equipment suppliers. The pictorial section, although not 
everything to be desired, adds flavor to the book and prevents it from becoming a 
mere compilation of articles and other information. Because of the international 
character of the publication, it would be helpful if a few notes were given concerning 
the contributors since it is not generally the case that authors are well known out- 
side their respective countries. 


Pavelle Color, Inc. 

533 W. 47 St., New York, N. Y. 

The High-Current Carbon Arc, by Wolfgang Finkelnburg 

Published (1947) by the Office of Military Government for Germany (U. S.) 
Field Information Agency, Technical, Final Report No. 1052, through the Office 
of Technical Services, U. S. Department of Commerce (Publication Board No. 
81644). Paper covers, photo offset from typewritten manuscript. 219 pages + x 
pages. 129 figures. 8 tables. 90 references. 7 1 /t X 10 inches. Price, $5.00. 

This book, in the first German edition of 1944, was prepared as a confidential 
text for the guidance of scientists in Germany working on carbon-arc searchlight 
development. Since the Allies came to rely on radar communication while the 
enemy was still expanding the size and intensity of his antiaircraft searchlights, 
these latter reached a much higher state of development than here in the United 
States. For instance, 450- and 1000-ampere carbon-arc searchlights were in 
active combat use and an advanced stage of development, respectively, in Ger- 
many* as compared with the 195-ampere maximum employed by United States 
forces. Since searchlight arcs differ in no important theoretical respect from those 
employed in motion picture photography and projection, these developments are 
of particular interest to technicians in the motion picture industry. The book 
presents "the whole knowledge" in Germany of the physical properties, theory, 
and application of carbon-arc light sources. In addition to a treatment of pre- 
viously published material, a large amount of hitherto unpublished information 
from the author's own laboratory as well as from other German workers and firms 
is included. 

Book Reviews 

The author classifies carbon arcs as of two fundamental types, low-current and 
high-current, depending upon the current density at the anode, and independent 
of the composition. He finds, for instance, that the plain "low-intensity" carbon 
arc, as we term it, behaves much as a high-intensity arc if it is operated at com- 
parable current densities. Of course, the unsteadiness and the noise of this over- 
load low-intensity arc are too great for most commercial uses, but the physical 
processes in the arc itself, particularly at the anode, are found to be identical with 
those which govern high-intensity arc operation. The first part of the book is 
devoted to a detailed description of the operating characteristics which distin- 
guish the two fundamental types of arcs. 

The next chapters deal with the operating properties of high-current carbon 
arcs, including not only the radiant output, but studies of arc temperature, carbon 
consumption, the transport of material through the arc, arc behavior in pure 
gases and at different pressures, magnetic properties of several types of arcs from 
the standpoint of stabilization, and chemical processes within the arc. This pre- 
pares the reader for the following theoretical section of the book, where a theory 
is presented explaining the fundamental differences between low- and high-current 
carbon arcs. In particular, the increase in arc voltage with increasing current 
which distinguishes the behavior of the high-current carbon arc from the negative 
characteristic of the low-current type, and the observed relations between current, 
brilliancy, and core composition are explained in terms of the author's theory. 

The concluding chapter of the book deals briefly with applications of the high- 
current carbon arc, the section on searchlights being of particular interest on 
account of the very high current units employed in Germany. The German 
practices in the motton picture studio and projection fields and medical thera- 
peutics are interestingly described. A concluding section, treating the carbon arc 
as a tool in high-temperature chemistry studies, well illustrates the author's hope 
that the book serve primarily as a basis and incentive for further research on the 
theory and application of the high-current carbon arc. To anyone interested in 
this field, the book provides a very worth-while background. 


National Carbon Company 

Cleveland 1, Ohio 

George Mitchell Receives ASC Award 

The American Society of Cinema- The award, the first given by the 

tographers presented George Mitchell, Society in its thirty years of existence, 

Associate member of the SMPE, with a was made in a surprise presentation to 

certificate of appreciation in recog- Mr. Mitchell, September 11, 1948. 

nition of his ceaseless pioneering in the Similar awards will be made by the 

field of motion picture photographic Society from time to time, to others 

equipment, and his immeasurable con- whose contribution to cinematography 

tribution to the advancement of cine- is considered equally noteworthy, 
matography as an art and as a science. 


Current Literature 

rriHE EDITORS present for convenient reference a list of articles dealing with 
J- subjects cognate to motion picture engineering published in a number of se- 
lected journals. Photostatic or microfilm copies of articles in magazines that are 
available may be obtained from The Library of Congress, Washington, D. C., or 
from the New York Public Library, New York, N. Y., at prevailing rates. 

American Cinematographer 

29, 11, November, 1948 
Cinecolor Moves Ahead (p. 373) 

Filming the Olympic Games (p. 374) 

Exponent of the Moving Camera 

(p. 376) H. A. LIGHTMAN 
Seven New Lenses for 16-Mm 

Cameras (p. 384) A. ROWAN 
Audio Engineering 

32, 11, November, 1948 
Sound on Film (p. 24) J. A. MAURER 

32, 12, December, 1948 
Suggested Wiring Standards for 
Motion Picture Recording Equip- 
ment (p. 16) G. RUDOLF 
British Kinematography 

13, 3, September, 1948 
The Origin and Development of the 
Matte Shot Process (p. 74) 
W. P. DAY 

Power Supplies for Motion Picture 
Studios (p. 83) F. S. HAWKINS 

13, 4, October, 1948 
The Processing of Colour Films 

(p. 109) J. H. JACOBS 
Safety Regulations in Kinemas and 

Theatres (p. 121) A. F. STEEL 
Light Production From the Carbon 
Arc (p. 128) H. P. WOODS 


28, 11, November, 1948 
High Fidelity Tape Recording (p. 16) 

International Photographer 

20, 10, October, 1948 
Thomascolor (p. 7) A. WYCKOFF 
Production in Germany (p. 12) W. B. 

International Projectionist 

23, 10, October, 1948 
Causes and Prevention of Damage 
to 35-Mm Theatre Release Prints 
(P- 5) 

Television: How It Works (p. 12) 

23, 11, November, 1948 
Safety Film: Projection Factors 

(p. 9) H. B. SELLWOOD 
Television: How it Works. Pt. 5 

(p. 17) W. BOUIE 
Proceedings of the I.R.E. 

36, 10, October, 1948 
The Chemistry of High-Speed Elec- 
trolytic Facsimile Recording 
(p. 1224) H. G. GREIG 

Radio and Television News 

40, 5, November, 1948 

The Recording and Reproduction of 

Sound. Pt. 21 (p. 50) O. READ 


New Products 

TPurther information concerning the material described below can 
JL be obtained by writing direct to the manufacturers. As in the case 
of technical papers, publication of these news items does not consti- 
tute endorsement of the manufacturer's statements nor of his products. 


Recogram Recorders Company, 
11338 Burbank Boulevard, North 
Hollywood, California, recently an- 
nounced the "Magnagram" M-116. 
This is a completely portable location 
recorder which is contained in two 
separate cases. 

Conting'ent upon film speed, the unit 
is capable of recording up to thirty 
minutes. With 400-foot reels the 
mechanical unit can be blimped for 
silent operation. The transparent door 
leaves the working parts readily visible. 

The magnetic film is caused to pass 
over an independent floating drum 

connected with a dynamically balanced 
flywheel in the retreat from the record- 
ing-reproducing heads. This motion, 
coupled with dampening arms confines 
flutter to approximately Vio of one per 

For ease in editing magnetic film the 
M-116 is designed to pass the track 
backwards as well as forwards over the 
heads without altering the normal 

All components of the magnagram 
are mounted on 19-inch Western Elec- 
tric relay-rack panels. The drive as- 
sembly employs a hysteresis-synchro- 
nous motor with prelapped nylon gears. 


New Products 

Further information concerning the material described below can 
be obtained by writing direct to the manufacturers. As hi the case 
of technical papers, publication of these news items does not consti- 
tute endorsement of the manufacturer's statements nor of his products. 

Photocell Cable 

Transradio, Ltd., 138A Cromwell 
Road, London, S.W.7, England, has 
announced a special photocell cable, 
type PC1-T. It is claimed that its 
losses are but a fraction of those com- 
mon to the usual type of low-loss cable. 
It is vibrationproof, nonmicrophonic, 
highly flexible, monaging, and un- 
p-fected by oils. 

Fig. 1 Photocell cable. 

This cable is designed to afford that 
constancy of electrical characteristics 
required for cable connections to port- 
able equipment. It finds particular 
application to all kinds of photoelec- 
tric-cell work, sound-film projection, 
and microphones, where stability of 
electrical values is necessary. 

Movette Camera 

American Cinefoto Corporation, 1560 
Broadway, New York 19, N. Y., re- 
cently introduced the Movette, a 
process for taking animated portraits. 

A high-intensity stroboscopic light 
combined with a multiple camera 
operates at 180 frames per minute and 
uses 4- X 5-inch cut film. 

The camera is without shutters. A 
built-in commutator arrangement fires 
the flash lamp in rapid sequences dur- 
ing which period the negative holder 
inside the camera completes its travel 
across the optical unit. The motor 
and flash cease to operate after the last 
picture has been taken. 

The Movette plugs in any 110-volt, 
60-cycle alternating-current light 
socket; the charging voltage is 2000 
volts and the effective flash of a 

A series of 42 pictures are produced 
in about 15 seconds. Developing the 
film and making an 11- X 14-inch 
enlargement are done in the conven- 
tional manner. 


Situation wanted in Television Broadcasting Industry. Possess 
_ 7 years' administrative, 5 years' motion picture engineering, and 4 
years' electronic teaching experience. For details write "Anony- 
mous," 190 Hutchison Blvd., Mount Veraon, N. Y. 

Mechanical and Electrical Design Engineer with experience on 
sound reproducers, amplifiers, and photoelectric devices. Kjeld 
Bogedam, c/o Nielsen, 244 Fisher Ave., Tottenville, S. I., N. Y. 


Journal of the 

Society of Motion Picture Engineers 



Three-Color Subtractive Photography 


Masking: A Technique for Improving the Quality of Color 

Reproductions T. H. MILLER 133 

Inter-Society Color Council Symposium Foreword 

C. R. KEITH 156 

Spectral Characteristics of Light Sources 


Color-Order Systems CARL E. Foss 184 

System in Color Preferences J. P. GUILFORD 197 

16-Mm Release Printing Using 35- and 32-Mm Film 


Three Proposed American Standards 223 

Joseph H. McNabb 231 

65th Semiannual Convention 232 

Motion Picture Test Films 234 

Book Review: 

"An Introduction to Color," by Ralph M. Evans 

Reviewed by Herbert T. Kalmus 236 

1949 Nominations 238 

Optical Society Meeting 239 

Current Literature 239 

New Products.. 240 


Chairman Editor Chairman 

Board of Editors Papers Committee 

Subscription to nonmembers, $10.00 per annum; to members, $6.25 per annum, included in 
their annual membership dues; single copies, $1.25. Order from the Society's General Office. 
A discount of ten per cent is allowed to accredited agencies on orders for subscriptions and 
single copies. Published monthly at Easton, Pa., by the Society of Motion Picture Engineers, 
Inc. Publication Office, 20th & Northampton Sts., Easton, Pa. General and Editorial Office, 
342 Madison Ave., New York 17, N. Y. Entered as second-class matter January 15, 1930, 
at the Post Office at Easton, Pa., under the Act of March 3, 1879. 

Copyright, 1949, by the Society of Motion Picture Engineers, Inc. Permission to republish 
material from the JOURNAL must be obtained in writing from the General Office of the Society. 
Copyright under International Copyright Convention and Pan-American Convention. The 
Society is not responsible for statements of authors or contributors. 

Society of 

Motion Picture Engineers 

342 MADISON AVENUE NEW YORK 17, N. Y TEL. Mu 2-2185 


Earl I. Sponable 
460 W. 54 St. 
New York 19, N. Y. 

Clyde R. Keith 
120 Broadway 
New York 5, N. Y. 


Loren L. Ryder 
5451 Marathon St. 
Hollywood 38, Calif. 

Peter Mole 

941 N. Sycamore Ave. 
Hollywood 38, Calif. 



VICE-PRESIDENT Ralph B. Austrian 

John A. Maurer 1270 Avenue of The 

37-01 31 St. Americas 

Long Island City 1, N. Y. New York 20, N. Y. 

Robert M. Corbin 
343 State St. 
Rochester 4, N. Y. 

William C. Kunzmann 
Box 6087 
Cleveland 1, Ohio 


David B. Joy 
30 E. 42 St. 
New York 17, N. Y. 



Alan W. Cook 
25 Thorpe St. 
Binghamton, N. Y. 

Lloyd T. Goldsmith 
Warner Brothers 
Burbank, Calif. 

Paul J. Larsen 
508 S. Tulane St. 
Albuquerque, N. M. 

Gordon E. Sawyer 
857 N. Martel Ave. 
Hollywood 46, Calif. 


James Frank, Jr. 
426 Luckie St., N. W. 
Atlanta, Ga. 

William H. Rivers 
342 Madison Ave. 
New York 17, N. Y. 

Sidney P. Solow 
959 Seward St. 
HoUywood 38, Calif. 

R. T. Van Niman 
4431 W. Lake St. 
Chicago 24, 111. 


Herbert Barnett 
Manville Lane 
Pleasantville, N. Y. 

Fred T. Bowditch 
Box 6087 
Cleveland 1, Ohio 

Kenneth F. Morgan 
6601 Romaine St. 
Los Angeles 38, Calif. 

Norwood L. Simmons, Jr. 
6706 Santa Monica Blvd. 
Hollywood 38, Calif. 

Three -Color 
Subtractive Photography 



Summary The color-vision characteristics of the eye are discussed and the 
rules which are followed are used to show the requirements for the "perfect" 
additive and subtractive three-color photographic processes. Since these re- 
quirements are not achieved in practice, a theoretical study of a practical 
color process may not always give an adequate analysis of its usefulness. 
However, such an analysis may point out some of the pitfalls which occur in 
practice. For example, many subjects may appear to be the same color to the 
normal eye and yet give different results when photographed. Also, any 
given color may be reproduced incorrectly by any process in use. 

The effects on picture quality of changes in contrast, balance, and a variety of 
other variables are shown. The restrictions which some of these factors 
place on the use of color films are mentioned. 

COLOR PHOTOGRAPHY is now an accepted reality and from all 
appearances is here to stay. During the past twenty years the 
motion picture industry has witnessed the slow but steady growth 
of color, both in the professional and amateur fields. Many prob- 
lems have been encountered and many, but not all of them, have been 
overcome. However, the average motion picture goer will lay his 
money down for a color picture in preference to one in black and white. 

How did color photography get here? Contrary to popular 
conception, color photography is not an invention and even the 
individual color systems and color processes are something more than 

The basic concepts of modern color photography are almost one 
hundred years old, and it is now a very complex and abundant field. 
A good many physicists, chemists, psychologists, physiologists, and 
artists have contributed. Until recently the terminology has varied 
with the profession as well as the eccentricity of the individual, so that 
much of the literature is not easy reading. 

Practical color processes do not emerge from theoretical discoveries 
alone. The successful existing color processes are here because of the 
parallel evolution of dye chemistry and photographic chemistry and, 

* Presented May 18, 1948, at the SMPE Convention in Santa Monica. 





most important, because of countless trial-and-error experiments, and 
the ingenious solution of thousands of small problems. 

Color photography must always start and end with the mechanism 
of color perception by the human eye, that is, the color process must 
first "see" the scene in a manner approximating that of the human eye. 
It must then reproduce that scene in such a manner that it seems 
plausible to the eye. 

It has long been known that all colors could be matched by mixing 
amounts of three so-called primary colors. With a given set of three 
primaries taken from the spectrum, each of the other colors of the 
spectrum can be duplicated by a mixture of certain intensities of the 
original three. There is an important reservation in this generaliza- 




Color-Mixture Curves 
of Spectrum 
Spectral Primaries- 
460, 530, and 650 





























s ^ 





>0 440 480 520 560 600 640 680 

Wavelength in Millimicrons 

Fig. 1 Color-mixture curves of the spectrum. 

tion, for it will be seen from Fig. 1 that in certain regions of the 
spectrum negative amounts of the three primaries must be permitted. 
Since negative quantities of the primaries cannot exist, the equivalent 
is obtained in practice by adding the third primary to the colors which 
cannot be matched. 

If another set of three wavelengths had been chosen as the pri- 
maries, a similar but different set of curves would have resulted. 

It is frequently desired to express the color-mixture data obtained 
with one set of primaries in the equivalent amounts of a different set 
of primaries. This process is illustrated in Fig. 2. 

The orange-yellow color at the top can be matched by the pri- 
maries, R, G, and B. The squares show the unit amounts of these 
primaries and the rectangular arrangement at the right shows the 




amounts of the three required to match the color. If we wish to 
express this color in terms of another set of primaries, R' ', G', and B', 
we can first find the amounts of R', G' , and B f which are required to 
match exactly the original unit amounts of J?, G, and B. By using 
R', G r , and B' in these ratios, the total amounts of R, G, and B shown 
in the upper right can be matched. Then the sum of the three values 
of R', the sum of the three values of G 1 ', and the sum of the three values 
of B' will match the original color. 

Fig. 2 Graphical transformation of primaries: 

Transformation of Primaries 
C = rR + gG + bB 
R = diiR -f- ai 2 (jr -\- CL\zB 

B : 

G = r(ttnR 4" d\yjr -f- (ii%B ) -(- g(a 2 \R -f- az 2 G' 

b(a u R f 

C ' = (ma 

(ra n 

This operation is usually carried out mathematically. The equa- 
tions which are involved are shown below. There a color, C, is 
matched by an amount, r, of the primary, R, an amount, g, of the 
primary, G, and an amount, 6, of the primary, B. These primaries 
can, in turn, be defined in terms of the amounts (an, ai 2 , a i3 , 021, etc.) 
of a second set of primaries, R', G 1 ', and B' '. If the unit amounts of 
R, G y B, and R', G', B' are defined by the amounts required to produce 
a white of a certain brightness, then the substitutions and factoring 
give the last form of the equation. 

In a similar manner, the amounts of the primaries, R', G', and B', 




which are required to match the various spectrum colors can be cal- 
culated from the color-mixture curves of the primaries, R, G, and B. 
This then gives us a new set of color-mixture curves. The change 
from one set to the other is a linear transformation. There are an 
infinite number of primaries and corresponding color-mixture curves 
which describe the color-vision characteristics of the human eye, and 
all of them tell exactly the same story. These of course include as 
primaries purely hypothetical colors which cannot exist in practice. 
Proper choice of hypothetical primaries can lead to mixture curves 
with no negative regions. 

Another important characteristic of color vision is the relative 
brightness of different colors. Equal energies of different colors are 


5 eo 



Luminosity Curve 

for a 
Normal Eye 


"400 450 500 550 600 650 700 

Wavelength in Millimicrons 

Fig. 3 Luminosity curve for a normal eye. 

not of equal brightness or luminosity. If the relative luminosity of 
spectrum colors of an equal energy is measured, the curve in Fig. 3 

For many years there was a good deal of confusion in the field of 
color measurement because a variety of workers used different pri- 
maries in determining the color-mixture curves and different methods 
of measuring the luminosity of the different spectrum colors. Differ- 
ent types of equipment led to slightly different results, because of 
certain inaccuracies and also because of the fact that the individual 
observers vary in their characteristics. For this reason a standard 
system of color specification became necessary for all the various 
workers in the field of color. 

This standard system was set up by the International Commission 




on Illumination and is called the ICI system. This group selected 
the previously adopted luminosity curve as a standard and defined its 
three primaries to meet certain requirements. First, all real colors 
should be matched with positive amounts. Second, one primary 
should be such that one of the mixture curves would be identical to the 
luminosity function. By using the best available color-mixture data, 
the primaries, X, F, Z, and the corresponding; mixture curves were 
established to define the "standard observer." Obviously the pri- 
maries do not represent real colors. These standard color-mixture 
curves are shown in Fig. 4. 







500 600 

Wovelength in Millimicrons 

Fig. 4. Standard ICI color-mixture curves. 

Maxwell, in 1855, suggested that positives made from black-and- 
white negatives which, in turn, were made through red, green, and 
violet filters, could be used to control the amounts of red, green, and 
violet light transmitted by filters and that these, when superimposed, 
would give a color reproduction of the original scene. 

All additive systems of color photography are modifications or 
applications of this invention made more than 90 years ago. 

This system of Maxwell's was an additive method. Shortly after- 
wards this principle was. extended by du Hauron, who showed that 
images made through the red, green, and blue filters and printed in 
superposition in cyan, magenta, \and yellow dyes or pigments would 
also give a fair reproduction of the original scene. This extension 


of Maxwell's system is the basis for all the subtractive color systems. 
However, it was many, many years before the sensitizers, dyes, and 
photographic materials in general were available for the application 
of these simple principles. 

For many years there were heated arguments as to the exact 
requirements for the sensitivity distributions of the three emulsion- 
filter combinations to be used in obtaining the three records for color 
photography. Certain workers in the field felt that narrow bands 
of sensitivity in the red, green, and blue regions of the spectrum gave 
the most satisfactory results. Others felt that the sensitivity dis- 
tributions of the three emulsions had to match the sensation curves 
of the eye. Others attempted to match the sensitivity-distribution 
curves of the emulsions with the absorption curves of the dyes or 
pigments being used in making subtractive prints. Although these 
earlier efforts did not lead to a resolution of these theoretical prob- 
lems, enough practical experience was gained so that when improved 
sensitizers, emulsions, and techniques of making colored photographic 
images were developed it was possible to work out empirically 
methods of making quite satisfactory color photographs. Continued 
progress in the "techniques" of making colored images and super- 
imposing them have brought color photography to the present state. 

Finally, knowledge as to the theoretical requirements for "exact" 
photographic color reproduction followed close on the heels of better 
data describing the color-vision characteristics of the eye. Hardy 
and Wurzburg 1 applied the principles of colorimetry and the char- 
acteristics of the human eye to the problem of establishing the 
theoretical requirements for the perfect additive three-color photo- 
graphic process They showed that the sensitivity distributions of 
the three emulsions used in obtaining the three images must corre- 
spond with the color-mixture curves determined with the three pri- 
maries which were used in showing the additive color picture. These 
distribution curves would be some linear transformation of the color- 
mixture curves obtained by using other primaries, including the 
standards selected by the ICI. Of course, for any primaries which 
could be used in practice, that is, real colors, even spectrum colors, 
these film sensitivities would require negative proportions of certain 
regions of the spectrum. Although a number of suggestions have 
been made as to how such negative sensitivities might be achieved, 
and some methods have been patented, to date no satisfactory prac- 
tical solution has been found. 


Later, Yule 2 and MacAdam 3 extended these principles to the 
problem of subtractive color photography. Although it was not 
possible to establish the so-called ideal dyes for use in subtractive 
photography, MacAdam showed that for one set of dyes actually 
being used in practice, it was possible to establish so-called additive 
primaries which would describe the behavior of subtractive mixtures 
of these dyes and was able to show that by the use of six masks it 
should be possible by photographic means to obtain a very close 
approximation to "exact" color reproduction. The basic principles 
were those developed by Hardy and Wurzburg, and the sensitivity 
requirements of the three emulsions in this process were the color- 
mixture curves derived from the primaries. These, of course, con- 
tained negative portions at certain regions in the spectrum. 

This can be summarized by stating that the theoretical require- 
ments which a subtractive color process must fulfill in order to give 
"exact" color reproduction have not been established completely. 
They do indicate a need for negative sensitivities and for the use of 
six masks. The first of these needs cannot be fulfilled at all and the 
second is entirely impractical. So, the color processes have to 
struggle along without fulfilling these requirements, and they do give 
satisfactory results. 

However, even though present-day color processes do give satis- 
factory results, there are certain deficiencies which must exist because 
of the failure to meet the requirement that the film-sensitivity dis- 
tribution be a linear transformation of the color-mixture data of the 
eye. For example, there are an infinite number of energy distribu- 
tions of light which appear the same to the eye. A color film will not 
necessarily see such colors as being alike. 

A pair of dye combinations which produce a very close visual match 
is shown in Fig. 5. These spectrophotometric measurements show 
the densities of the two combinations to light of the various wave- 
lengths in the visible range. As is often the case, these colors which 
appear to be identical have very different absorption characteristics. 
When photographed with one of the commercially successful color 
films, the resulting photographs are also very different. 

At first thought, one may say, "This doesn't make too much differ- 
ence because we shall never encounter two colors of this sort side by 
side." However, the fact that the two colors do not match indicates 
that at least one of them is not properly reproduced. In fact, any 
color might be improperly reproduced by any of the present color 




processes which in normal practice give excellent results. For- 
tunately most of the colors which we normally encounter have more 
or less continuous light-absorption ' bands and are reproduced fairly 
accurately. It may appear that an undue amount of emphasis is 
being put on this type of problem. The important point is that in 
dealing with flowers, new types of fabrics, or with new color situations 
in general, it is wise to make a test with a given photographic process 
to see that it will reproduce adequately the specific colors which are 
important rather than to assume that the process is perfect and start 

i i.o 


8 0.8 




500 600 

Wavelength in Millimicrons 

Fig. 5 Spectrophotometric curves of matching colors. 

For most practical applications of color photography, the repro- 
duction of colors need not be theoretically perfect. Even with very 
pleasing color pictures, an analysis of the individual colors will reveal 
considerable departure from the hue, saturation, and brightness of the 
original colors of the scene. However, when combined in a picture 
of familiar and pleasing composition, the color reproduction is plau- 
sible enough to give the impression of correct reproduction. 

Let us now look at some of the requirements which can and must be 
fulfilled in obtaining satisfactory color photographs. 

The first of these requirements is color balance. This is usually 
best observed in the accuracy with which grays of various bright- 
nesses are reproduced. One might consider this the minimum 
requirement of a color process. However, the errors encountered 




in matching grays to the original subject are present in about the same 
degree in the reproduction of all colors. In the case of pastels and 
other colors of low saturation, this error in balance may become a 
serious distortion. 

Color balance is measured by reading a scale of grays with a color 
densitometer and plotting the densities of the dyes against the 
logarithm of the exposure. By definition, 4 ' 6 the equivalent neutral 
densities (END) of the dyes of a given color process are those which, 
in superposition, will appear gray under the viewing conditions for 
which the color film is designed. A correctly balanced color process 




Log Exposure 




Fig. 6 Correctly balanced neutral scale. 

would have a gray scale in which all three dye curves were super- 
imposed. Slight deviations from this ideal are usually encountered 
at very low densities and also in the region of maximum density. 

If, however, the color balance is uniformly high in any of the dyes, 
magenta in the case of Fig. 7, the picture through such a process will 
also show a decided shift to a magenta balance. This is noticed not 
only in the reproduction of grays but in a change in all the colors of 
the pictures. Thus, blues become more purple, yellows become more 
orange, and greens become darker. This type of distortion results 
also from any change in color temperature of the exposing light from 
that for which the color film has been balanced. 

This uniform shift in color balance of a color film may at first 




appear to be very objectionable but when a picture is viewed by pro- 
jection in a darkened room the distortion appears to become less ob- 
jectionable with continued viewing. This accommodation of the 
visual process, or color adaptation, does tend to make an off-balance 
picture appear more nearly satisfactory by projection. If, however, 
as in motion picture projection, the color balance shifts from scene to 
scene the color change is very noticeable. 

In Figs. 6 and 7 the gammas of all the dye scales were equal. 
If the gammas of the three dye images are not equal, Fig. 8, the color 
photograph varies in color balance from one density level to another 




Log Exposure 




Fig. 7 Magenta-balanced neutral scale. 

and the distortion in color reproduction varies depending on the color 
and its brightness. In a picture through the process represented 
in Fig. 8, the light densities would be much too green and the darker 
portions of the pictures much too magenta. This is a very unde- 
sirable type of distortion, for the eye cannot become adapted to both 
errors. Such a picture continues to be objectionable, no matter how 
long we look at it. From the shapes of the curves it is easy to see why 
such a distortion has been termed a kink. 

Assuming that the first requirement is fulfilled and that correct 
color balance and matched relative gammas of the three dye images 
can be achieved, and these are no small assumptions, we are still faced 




with a difficult decision: "What gamma or contrast level is most 
desirable for a specific color process?" In black-and-white photog- 
raphy, the question is completely answered by the requirements for 
pleasing tone reproduction. In color photography, the problem is 
complicated by the fact that color saturation varies with the gamma. 
At low gamma, we can have all the advantages of pleasing tone ren- 
dition and greater latitude but we must pay for these advantages by 
sacrificing color saturation. At high gamma, color saturation is 
satisfactory but latitude must be low. 




Log Exposure 




Fig. 8 Neutral scale of unmatched gammas. 

The relationship between gamma and color saturation can be ex- 
plained quite easily by expressing the amounts of the three image 
dyes present in a given area of a color photograph in terms of equiva- 
lent neutral density. For example, suppose that a given color is 
reproduced by a color process having a gamma of 1.0 by the amounts 
of the dyes: cyan, 0.8; magenta, 0.1; and yellow, 0.6. If the same 
process were operated at a gamma of 1.5, the reproduction densities 
would be: cyan, 1.2; magenta, 0.15; and yellow, 0.9. It is possible 
to determine the color densities (in terms of END) over and above the 
gray content (in terms of END) merely by subtracting the density of 
the dye occurring in the lowest amount. The results of this sub- 
traction from the two groups of densities above are given in Table I . 



Cyan Magenta Yellow Cyan-Magenta Yellow-Magenta 

At 7 = 1-0 
At? = 1.5 






The much greater density differences at the gamma of 1.5 represent 
a significant increase in color saturation.* 

Experience has shown that the color saturation obtained at higher 
contrast is quite desirable so that in practice, processes are usually 
operated at a relatively high contrast. The desired tone repro- 
duction must be obtained by much natter lighting than normally 
would be used in black-and-white photography. It may be noted in 
passing that the opposite approach is not satisfactory, that is, a low- 
contrast process with contrasty lighting. The maximum color 
saturation possible is limited by the density range and gamma of the 
color process and is only slightly affected by lighting variations. 

The decision between usable contrast and acceptable color satura- 
tion again arises in duplicating a color film. The process of duplicat- 
ing a color film results in a loss in color saturation, providing the 
contrast is reproduced at the same level as in the original picture. 
This loss in saturation is a result of the properties of the dyes available 
for use in color photography. The only way this color-saturation 
loss can be improved is by increasing the contrast of the reproduction. 
This again results in a very definite compromise in the over-all quality 
of the reproduction. A more complete discussion of this problem is 
given in a separate paper by Miller. 6 

After considering some of the theoretical problems involved in 
obtaining more nearly perfect color reproduction and some of the more 
elementary variables in color processes, we should like to stress that 
what is necessary and what is desired in a color process depend very 
largely on the manner in which the process is to be used. The re- 
quirements for a good motion picture print in color are different in 
marly respects from the requirements for a good reflection print 

* This method of expressing colors in terms of END does not express a quantitative 
value for hue and saturation such as a Munsell notation or the like, but it is a 
useful technique in the field of color photography. 


Evans 7 has directed attention to the many psychological effects 
which complicate any orderly analysis of color vision and color 
photography. Brightness constancy, color and brightness adapta- 
tion, and simultaneous contrast have a profound influence in all color 

These phenomena result from the fact that the visual process does 
not function merely as a physical instrument for measuring the stimuli 
from different areas. On the contrary, the appearance of an object 
is always affected by the spatial relation of the object to other objects 
and to lighting conditions. The eye always sees things as the 
observer thinks they really are rather than as they happen to appear 
at the moment. A simple example is that of a white object in a 
shadow near a black object in full sunlight. Although the luminance 
of the white object may be much less than that of the black object 
under these conditions, the eye immediately recognizes the true 
brightness relationship of the two objects. This brightness-constancy 
effect may not be shown to the same degree in a photograph as in the 
original scene since the viewing conditions are in most cases entirely 

The eye varies in sensitivity to light over a considerable range 
depending on the intensity level under which it is used. Something 
similar to this brightness adaptation causes the eye to become adapted 
to colored light so that it tends to accept that color as white. The 
maximum effect, of course, is realized when all of the light reaching 
the eye is of the same color. 

Simultaneous contrast has been related to the adaptation of the eye 
to local areas of a picture. That may be simplified by stating that 
areas of complementary or contrasting colors appear to be increased 
in saturation by their proximity within the picture. In addition to 
selecting colors which produce a desired hue in the finished color 
photograph, simultaneous contrast can be used to striking advantage 
in obtaining pleasing color pictures. 

These effects are distinctly beneficial in the projection of color 
transparencies in a darkened room and are therefore very important 
in the success of many motion picture processes. Furthermore, 
these psychological effects explain in part why the data obtained in an 
isolated field of a colorimeter and the mathematical derivations from 
such data may have very little correlation with the infinite variety of 
conditions which can be encountered in photography and in the 
presentation of the resulting pictures of everyday objects. 


You will recall the emphasis on the word approximation, in com- 
paring the color photographic process to the visual process. We think 
it is still safe to state that the perfect color process has not yet been 
realized. There are, however, many successful color processes which 
can give very pleasing results in spite of the many compromises which 
must be present in each system. This means that to obtain satis- 
factory results the user must learn quite a lot about the color process 
with which he is working. As he learns what a particular color 
process will, and equally important, what it will not do, his success 
with that process will become more consistent. 

Close co-operation between the user of color photographic ma- 
terials and the manufacturer of them has been and will continue to be 
very important in obtaining satisfactory results with what we now 
have and know. This sort of co-operation is also necessary for the 
introduction of improved color photographic materials and techniques 
which will give better results. 


(1) A. C. Hardy and F. L. Wurzburg, Jr., "The theory of three-color reproduc- 
tions," J. Opt. Soc. Amer., vol. 27, pp. 227-240; July, 1937. 

(2) J. A. C. Yule, "The theory of subtractive color photography," /. Opt. Soc. 
Amer., vol., 30, pp. 322-331; August, 1940. 

(3) D. L. MacAdam, "Subtractive color mixture and color reproduction," J. 
Opt. Soc. Amer., vol. 28, pp. 466-480; December, 1938. 

(4) R. M. Evans, "A color densitometer for subtractive processes," J. Soc. 
Mot. Pict. Eng., vol. 31, pp. 194-202; August, 1938. 

(5) M. H. Sweet, "A precision direct-reading densitometer," J. Soc. Mot. Pict. 
Eng., vol. 38, pp. 148-173; February, 1942. 

(6) T. H. Miller, "Masking: A technique for improving the quality of color re- 
productions," J. Soc. Mot. Pict. Eng., this issue, pp. 133-155. 

(7) R. M. Evans, "Visual processes and color photography," J. Opt. Soc. Amer., 
vol. 33, pp. 579-614; November, 1943. 

Masking: A Technique for 
Improving the Quality 
of Color Reproductions* 



Summary Currently available subtractive color-photography processes 
provide pleasing pictures of most natural objects. However, when an original 
color photograph is the subject, as in the cases of duplicating and copying, the 
resulting reproduction is usually not satisfactory when compared with the 
original. Differences between the original and the reproduction are primarily 
due to the high photographic contrast and the optical characteristics of the 
dyes in the original. Masking to improve the quality of color reproductions 
involves making an auxiliary image, generally by a photographic method, and 
registering it with the original color transparency. Reproductions are made 
from the combined transparency and mask. 


BOTH THE SCIENCE and art of color photography have developed rap- 
idly since the introduction of Kodachrome film in 1935. Color 
motion pictures and still-picture transparencies of excellent quality 
are now a reality for both amateur and professional photographers 
using only conventional cameras and processing equipment. 

In the early years of Kodachrome the original color pictures were 
used only for projection for the enjoyment of the family and friends. 
In manufacturing and processing, therefore, the films were balanced 
for optimum quality when shown by projection in darkened rooms. 
Even now most color pictures are used only for projection and the 
processes are still balanced for best results for this condition. 

Interest in color and color pictures spread rapidly and with the 
introduction of Kodachrome and Ektachrome sheet films, professional 
and commercial color photography gained further momentum. 
Color added interest and appeal to educational, entertainment, indus- 
trial, and other motion pictures previously made only in black and 
white. Color slidefilms, color prints, and color reproductions of 
photographs in magazines and on billboards are now becoming the 
rule rather than the exception. 

* Presented May 18, 1948, at the SMPE Convention in Santa Monica. 


134 MILLER February 

Contrary to the amateur and early professional practice, a single 
original transparency or motion picture seldom satisfies the require- 
ments of the present-day professional color photographer. The pro- 
fessional generally must supply many reproductions from his original : 
duplicate motion pictures or slidefilms for simultaneous showings in 
all parts of the country, color prints for publicity releases, advertising 
folders, display ads, and many other forms of color reproductions. 

It is more difficult to make good color reproductions than it is to 
make good color originals. The nature of the problem (and some of 
the reasons) can be stated briefly as follows : 

(1) The original pictures made by three-color subtractive processes, 
such as Kodachrome, for example, provide very acceptable, though 
perhaps not exact, reproductions when used to photograph most nat- 
ural objects. There are some unusual objects with sharp absorption 
bands that do not photograph well because of the relationships be- 
tween the absorption bands of the object and the sensitivities of each 
of the three film-emulsion layers. A color photograph is one of these 
objects. The spectral characteristics of a color photograph and the 
subject are frequently quite different and while the two may appear 
similar they may photograph differently. 

(2) The photographic contrast of color films is relatively high. 
With the dyes that can be used in a practical color process this is 
necessary to provide sufficient color saturation. The contrast, though 
high, is not beyond the acceptable limit for making color originals. 
However, when the high-contrast original is reproduced on' a second 
high-contrast material, the contrast is increased and generally, then, 
exceeds the acceptable limit. 

(3) Color originals are seldom judged by comparing them with the 
subject. The quality of duplicates is almost always judged on the 
basis of side-by-side inspection or projection with the original color 

(4) The dyes that can be used in color processes are not perfect. 
Their hues and relative brightnesses are often far from ideal, though 
close enough to provide pleasing color originals. When a reproduction 
is made the optical characteristics of the dyes cause the exaggeration 
of the relative brightnesses of some colors and shifts in the hues of 
others, often to the extent that the reproduction is not acceptable. 

In order to use films of the Kodachrome type in the production of 
theater release prints it would be necessary to make a first repro- 
duction to work in the special effects. The release prints would then 

1949 MASKING 135 

be reproductions of reproductions and the deficiencies would be still 
more serious. 

Masking is a technique, used in making color reproductions, either 
to correct fully or to minimize the reproduction deficiencies resulting 
from the high photographic contrast or the optical characteristics of 
the dyes in the original photograph or both. 

A mask is an auxiliary image (generally a photographic image) used 
in register with an original color picture to modify the characteristics 
of the original for purposes of reproduction. Depending on the na- 
ture of the mask it may be used to modify contrast, to change the 
relative brightnesses of some colors, or to shift the hue or saturation of 
some colors. 

Masks are generally made by contact-printing the original color 
photograph onto a light-sensitive film, processing the film, and then 
visually or mechanically registering the color original and mask. 
The reproduction is made from this combination. 

Masks may be negative (if made from a positive), or positive (if 
made from a negative). Masks may be black and white (as in the 
case of silver masks) or colored (as in the case of single or multi- 
colored-dye masks). 

The combination of mask characteristics and the contrast required 
depends upon the nature of the original, the character of the unmasked 
reproduction, and the modifications desired. Since it is the purpose 
of this paper to describe masking for the improvement of color repro- 
ductions, the following will be considered in order: 

(1) the characteristics of the color original; 

(2) the characteristics of the unmasked color reproduction; 

(3) the modifications of the color original which are necessary if 
the reproduction is to match the original. 

Reproductions of color transparencies may be made on films incor- 
porating dye systems which are either the same or different from the 
dye system of the original film. To differentiate between these two 
techniques it is convenient to refer to the former as duplicating and to 
the latter as copying. 

In descriptions of color reproductions made on the same type of 
films as the original photographs the technique will be referred to as 
duplicating and the result, a duplicate. An example is the printing of 
a Kodachrome original onto a Kodachrome film. When the repro- 
duction is made by using a different process from the original, the 




technique will be referred to as copying and the result a copy. The 
masking problem differs for these two reproduction techniques. 


Since Kodachrome film is used for making more original color pic- 
tures than any other process, the characteristics of such originals will 
be discussed first. Other processes are essentially the same and the 
specific and important differences will be noted later. There are 
several different Kodachrome films with different dye characteristics ; 
so the examples which follow are not to be regarded as representing 
any specific Kodachrome film, but simply as representative of a re- 


400 500 600 700 

Fig. 1 Spectrophotometric curves of ideal and practical 

cyan dyes. 

versal color process where the couplers are added in the processing 
rather than incorporated in the film at the time of manufacture. Of 
the two reversal processes, Kodachrome is an example of the former 
type and Ektachrome is an example of the coupler-included type. 

Tungsten lamps used in projectors and illuminators radiate energy 
at all wavelengths and when viewed produce a sensation of white 
light. When the spectral distribution of energy of this light source is 
altered by the selective absorption of a dye, the sensation of white is 
no longer produced and the light appears to be colored. 

Both Kodachrome and Ektachrome pictures consist of three super- 
imposed dye images, one cyan, one magenta, and one yellow. The 
dye images in combination absorb the complementary colors of those 




seen in the color picture. The cyan dye absorbs its complementary, 

red, the magenta dye absorbs green, and the yellow dye absorbs blue. 

Absorption characteristics of dyes are shown by spectrophotometric 








Fig. 2 Spectrophotometric curves of ideal and practical 
magenta dyes. 







Fig. 3 Spectrophotometric curves of ideal and practical 
yellow dyes. 

curves, the extent of the absorption being indicated by the height of 
the curve at each wavelength. 

An ideal cyan dye would absorb red only as indicated by the dotted 
line in Fig. 1. The characteristic of the Kodachrome-type cyan, 

138 MILLER February 

which absorbs an appreciable amount of green and blue, is shown by 
the solid line. Since the green and blue absorptions are about equal, 
the dye by itself appears relatively darker than the ideal cyan indicated. 

An ideal magenta dye would absorb green only, transmitting all of 
the red and blue (Fig. 2). The Kodachrome-type magenta does not 
absorb all of the green and does absorb some blue and red. If the 
magenta absorbed blue and red equally, it would simply be darker 
than an ideal dye, but since it absorbs much more blue than red, the 
hue is different from the ideal magenta it is shifted toward red. 

An ideal yellow would absorb blue only (Fig. 3), transmitting all the 
green and red. The Kodachrome-type yellow does not absorb all 
blue, does absorb some green, and a very slight amount of red. 

To simplify the following descriptions, schematic spectrophoto- 
metric curves will be used to illustrate dye characteristics. These are 
shown in Fig. 4 (a), (b), and (c). The red absorptions of the magenta 
and yellow dyes are neglected because they are unimportant in the 
problem of reproduction as summarized in Table I. 


Gamma Values* 
Dyes Blue Green Red 

Cyan 0.45 0.45 1.5 

Magenta 0.45 0.90 

Yellow 0.60 0.15 

* Gamma values of cyan- , magenta- , and yellow-dye images resulting from ex- 
posure to white light and measurement with blue, green, and red light. These 
values and those given in subsequent tables are measured with respect to the orig- 
inal scene. 

Combined in equal quantities the Kodachrome-type dyes de- 
scribed form a neutral scale because the sums of the absorptions of 
red, green, and blue light by the three dyes are equal. The combined 
dye diagrams (Fig. 5) show the contributions of each dye to the 


By means of a single, simple formula it is possible to calculate the 
reproductions of the principal colors and their neutral scale in the 
duplicate. The principal colors referred to are the pure cyan, 




magenta, and yellow dyes of the original and their combinations in 
pairs to form red, green, and blue. For color duplicating, as for all 
other photographic reproductions, the contrast (or gamma) of the 
original multiplied by the gamma of the duplicating film gives the 
contrast of the duplicate picture. 

7 (original) X 7 (duplicating film)' = 7 (duplicate) 

B . G 





Fig. 4 Schematic spectrophotometric curves 
(a) cyan, (b) magenta, and (c) yellow dyes. 


When an original of this type is printed onto a film of the same 
kind, all the blue-light contrast is recorded by the blue-sensitive layer 
of the duplicating film and reproduced as yellow dye with a gamma of 
0.6. All green-light contrast is reproduced in the green-sensitive 

140 MILLER February 

layer as magenta dye (gamma 0.9), and all red-light contrast is repro- 
duced in the red-sensitive layer as cyan dye (gamma 1.5). 

The reproduction of the original cyan Applying the formula above 

gamma of original = 1.5 (cyan-dye gamma to red light) 
gamma of duplicating film = 1.5 (cyan-dye gamma to red light) 

gamma of duplicate = 2.25 (cyan-duplicate gamma to red light). 

Since the dye in the duplicate has the same chemical structure as 
the original cyan, its blue, green, and red absorptions have the same 
relationship as those of the original. Therefore, the cyan in the dupli- 
cate with red-light contrast of 2.25 has green-light contrast of 0.68 
and blue-light contrast of 0.68. 

B G R 

Fig. 5 Contributions of cyan, magenta, and yel- 
low dyes to neutral scale. 

Because the cyan dye in the original absorbs green (gamma = 0.45) 
duplication of a cyan image will result in a magenta image coincident 
with the cyan duplicate. The green-light contrast of the original 
cyan is 0.45. This is reproduced as magenta at a gamma of 0.9, hence 
the magenta addition has a green-light contrast of 0.45 X 0.9 = 0.41. 
The magenta given in the example also absorbs blue and the blue con- 
trast is 50 per cent of the green contrast. The blue-light contrast of 
this magenta addition is thus 0.21. 

The original cyan also absorbs blue (0.45), and this is reproduced 
as yellow dye at a gamma of 0.60. There is, therefore, a yellow-dye 
addition to the cyan image in the duplicate, the contrast of the yellow 
addition being 0.45 X 0.60 = 0.27. This yellow dye absorbs some 
green, 25 per cent of the blue absorption or 0.07. 

1949 MASKING 141 

The result of duplicating; the cyan image of the original is thus : 

Gammas to 
Blue Green Red 

0.68 0.68 2.25 cyan dye 

0.21 0.41 0.0 magenta-dye addition 

0.27 0.07 0.0 yellow-dye addition 

1.16 1.16 2.25 

A comparison of the original cyan and the duplicate from it (Fig. 6) 
shows the duplicate to be higher in contrast and density than the 


B G R 

Fig. 6 Schematic spectrophotometric curves 
comparing an original cyan scale with the duplicate 
made from it. 

original. It is darker but it has not changed hue because the green 
and blue absorptions are equal in the duplicate as they are in the origi- 
nal. While the high cyan contrast is due to the high contrast of the 
original and the duplicating film, the magenta- and yellow-dye addi- 
tions are due to the green and blue contrast of the original cyan. 
The reproduction of the original magenta 

gamma of original = 0.9 (magenta-dye gamma to green light) 
gamma of duplicating film = 0.9 (magenta-dye gamma to green light) 

gamma of duplicate = 0.81 (magenta-duplicate gamma to green light). 

In the example chosen the magenta dye has blue-light contrast 

142 MILLER February 

equal to 50 per cent of its green-light contrast, and since the green- 
light gamma of the magenta dye in the duplicate is 0.81 the blue-light 
contrast is 0.40. These are the optical characteristics of the middle 
layer of the duplicate. 

Because of the blue-light absorption of the magenta dye in the origi- 
nal a yellow-dye image is added to the magenta duplicate. The blue- 
light contrast of the original magenta is 0.45 and this is reproduced as 
yellow dye at a gamma of 0.6, giving a yellow addition with a blue- 
light contrast of 0.27 and a green-light contrast of 25 per cent of this 
or 0.07. 

Fig. 7 Schematic spectrophotometric curves 
comparing an original magenta scale with the 
duplicate made from it. 

The result of duplicating the original magenta image is thus: 

Gammas to 
Blue Green Red 

0.40 0.81 0.0 magenta dye 

0.27 0.07 0.0 yellow-dye addition 

0.67 0.88 0.0 

The blue absorption of the original magenta was 50 per cent of the 
green. This blue absorption resulted in the addition of yellow to the 
magenta areas in the duplicate raising the blue absorption to over 70 
per cent of the green (Fig. 7). This shifts the hue of the magenta 
toward the red, meaning that original magenta areas are more red in 
the duplicate. The green-light contrast is lower in the duplicate than 
in the original. This indicates a loss in the saturation of magenta 
areas in addition to the hue shift. 

The reproduction of the original yellow 

gamma of original = 0.6 (yellow-dye gamma to blue light) 
gamma of duplicating film = 0.6 (yellow-dye gamma to blue light) 

gamma of duplicate = 0.36 (yellow-duplicate gamma to blue light). 




The original yellow is relatively very low contrast. When dupli- 
cated the yellow has a gamma of only 0.36 to blue light and 25 per cent 
as much or 0.09 to green light. 

Since the original yellow dye absorbs some green light, magenta 
dye is added to the yellow areas in the duplicate. The green-light 
contrast of the original yellow is 0.15, and it is reproduced as magenta 
with a green-light gamma of 0.9 making a magenta image with green- 
light contrast of 0.14. This additional magenta absorbs blue with a 
contrast of 0.07 (50 per cent of 0.14). 

Results of duplicating the yellow are thus : 




Gammas to 






yellow dye 
magenta-dye addition 



Fig. 8 Schematic spectrophotometric curves 
comparing an original yellow scale with the 
duplicate made from it. 

A comparison of the yellow original and duplicate (Fig. 8) shows the 
contrast, hence the color saturation, of the duplicate considerably 
lower. Actually, it is so low that there is almost no yellow in the 
duplicate. Whereas the original yellow exhibited green absorption of 
only 25 per cent of the blue, the duplicated yellow has a green absorp- 
tion of over 50 per cent of the blue. This causes a shift in hue of the 
yellow toward orange. 

Red in an original transparency is produced by the combination of 
magenta and yellow dyes. Therefore, the reproduction of a red area 
can be calculated by adding the characteristics of the magenta and 
yellow reproductions. The same procedure can be applied to deter- 
mine the nature of a duplicate green (cyan plus yellow) and a dupli- 
cate blue (cyan plus magenta). 




Red in Original 
B G R 

0.45 0.90 0.0 magenta 
0.60 0.15 0.0 yellow 

1.05 1.05 0.0 

Red in Duplicate 
B G R 



0.0 from magenta 
0.0 from yellow 

Green in Original 
B G R 



1 . 5 cyan 
0.0 yellow 

1.05 0.60 1.5 

Blue in Original 
B G R 



1.5 cyan 
0.0 magenta 

0.90 1.35 1.5 

1.10 1.11 0.0 

Green in Duplicate 
B G R 

1.16 1.16 2.25 from cyan 
0.43 0.23 0.0 from yellow 

1.59 1.39 2.25 

Blue in Duplicate 
B G R 

1.16 1 . 16 2 . 25 from cyan 
0.67 0.88 0.0 from magenta 

1.83 2.04 2.25 

From all of the above calculations the effect of duplication on the 
principal colors can be tabulated as follows : 

cyan much darker 

magenta slightly desaturated, hue shifted toward red 

yellow greatly desaturated, hue shifted toward red 

red duplicate closely matches original 

green much darker, slight shift in hue toward blue 

blue much darker, shift in hue toward green 

Since the original neutral scale is a combination of all three dyes, 
its reproduction is the sum of the reproductions of cyan, magenta, 
and yellow. 

(duplicate from orginal cyan) 
(duplicate from original magenta) 
(duplicate from original yellow) 

Gammas to 
















This reproduction is neutral and its contrast is 2.25. This figure 
checks with that obtained in another way of calculating the neutral 
gamma. The original contrast is 1.5 and the duplicating film has 
contrast equal to 1.5; then the duplicate has contrast 2.25. 

In summarizing the characteristics of the unmasked reproduction, 

1949 MASKING 145 

it is important to repeat that the reproduction deficiencies are pro- 
duced by the optical characteristics of the dyes in the original color 
transparency. In understanding color-reproduction problems, it is 
important to remember that these are characteristics of the dyes them- 
selves and are not the results of improper film manufacturing or 

It is difficult to predict whether or not original color pictures would 
appear better than those afforded by present processes if "ideal" dyes 
were available. The real advantage of the "ideal" dyes would be 
realized in duplicating as there would be no "additions" to the dupli- 
cate of each layer as now encountered because dyes absorb in regions 
other than those for which they are primarily intended. 


It is impossible to* single out one of the four undesirable dye absorp- 
tions described as the most serious of all, for each unwanted absorption 
is effective twice: once directly and once indirectly. The direct 
effects, additions of unwanted dye to each image, have been de- 
scribed. In addition, the presence of an unwanted absorption in one 
dye means that the dye primarily intended to absorb that color must 
be lower in contrast to maintain a satisfactory neutral scale. The 
color of the primary absorber in this relationship is thus desaturated 
in the original picture and even to a greater extent in the duplicate. 
The yellow dye of the example illustrates this effect. Since the cyan 
and magenta each absorb an appreciable amount of blue, the yellow 
must be quite low in contrast. 

In a reproduction of a skyscape or a seascape, the direct effects of 
the blue and green absorptions of the cyan would certainly spoil the 
duplicate. In a reproduction of a yellow dress or flowers, the green 
absorption of the yellow dye would be the only direct source of trouble, 
though the indirect effect of the cyan-dye characteristics on the yellow 
saturation must not be forgotten as they may be just as important. 

In transparencies containing all colors the most objectionable repro- 
duction errors are frequently the excessive darkening of cyans, blues, 
and greens caused by the characteristics of the original cyan dye. 

The green and blue absorptions of the original cyan of the example 
have gammas of 0.45 which are positive since the image is positive. 
A negative silver mask of the cyan layer alone, developed to a con- 
trast of 0.45 and bound with the original cyan, would subtract 0.45 
from the contrast of the cyan to blue, green, and red light as follows : 

146 MILLER February 

Gammas to 
Blue Green Red 

0.45 0.45 1.5 unmasked cyan-dye image 

-0.45 -0.45 -0.45 mask image 

0.0 0.0 1.05 

In practice such a mask can be made easily by printing the Koda- 
chrome or similar color transparency onto black-and-white film using 
red light. The cyan dye is the only dye that absorbs red, and thus the 
black-and-white picture is a record of the cyan layer only. The red- 
light contrast of the cyan dye in the original is 1.5, and the contrast 
necessary in the mask for canceling the contrast to green and blue is 
0.45. Therefore, 1.5 X gamma to which the mask must be developed 
= 0.45. The mask development is thus = 0.3. 

This is a red-light mask. It is a 30 per cent mask as far as the red- 
light contrast of the cyan is concerned but a wliole or 100 per cent 
mask as far as the blue and green contrasts are concerned. The mask, 
of course, covers the neutral scale as well as the picture, reducing the 
neutral-scale contrast from 1.5 to 1.05. 

While the red-light mask cancels the blue- and green-light contrast 
of the original cyan, lowers the red-light contrast of the cyan, and 
lowers the neutral-scale contrast, it does not alter the characteristics 
of the magenta or yellow images. Magenta and yellow areas in the 
original transmit all the red light used to expose the mask, and the 
contrasts of these two layers are unchanged by the mask. Only the 
density of the magenta and yellow areas is increased. 

With the mask in register with the original transparency, the combi- 
nation is characterized as follows : 

Gammas to 
Blue Green Red 

cyan dye plus mask 0.0 0.0 1.05 

magenta dye 0.45 0.90 ... 

yellow dye 0.60 0.15 ...' 

1.05 1.05 1.05 

The Duplicate from the Masked Original 

The red-light contrast of the cyan dye with the mask over it is 1.05. 
The gamma of the duplicating cyan to red light is 1.50. Thus the 
duplicate from the masked cyan has red-light gamma = 1.575 or ap- 
proximately 1.50. As in the original, the cyan dye in the duplicate 
absorbs blue and green light with contrasts of 0.45 each. 

1949 MASKING 147 

There are no additions of magenta or yellow since the masked cyan 
has no green or blue contrast. The reproduction of the masked cyan, 
therefore, almost exactly matches the original cyan. The magenta 
and yellow reproduce exactly as they did when no mask was used, 
for the mask does not influence the contrast of either of these original 

When the duplicates from the unmasked and masked cyans are 

compared, the duplicate from the masked cyan shows improved 

brightness but no change in hue (Fig. 9) . Since blue and green in the 

. original contain much cyan, the red-light mask results in improved 


G R B G R 

Fig. 9 Schematic spectrophotometric curves 
comparing duplicates from unmasked and masked 
cyan scales. 

brightness notably in cyans but also in blue and green reproductions. 
In practice the use of a red-light mask also improves the reds. With- 
out a mask, since the cyan is darker in the reproduction, the tendency 
is to print the duplicate somewhat lighter over-all to prevent exces- 
sively dark skies, grass, and other combinations of which cyan is a 
part. Then the reds are too light. When the mask is used and the 
duplicate properly printed, the reds reproduce relatively darker than 
in the unmasked duplicate. 

The red-light mask does not compensate for any hue shifts en- 
countered during duplicating. It only improves the relative bright- 
nesses of cyans, blues, and greens compared with reds, yellows, etc., 
and is often referred to as a relative-brightness mask. 

148 MILLER February 

The masked neutral scale has a contrast of 1.05 and the duplicating 
film has a gamma of 1.5; so the contrast of the neutral scale in the 
reproduction is approximately 1.5, the same as in the original. 


The use of a single silver mask as described above represents the 
most practical masking technique in duplicating, and recognition of the 
following features of the single silver mask is important to its suc- 
cessful use. 

(1) With a single silver mask in register with a color transparency 
for duplication, the neutral scale of the transparency remains neutral 
but is lower in contrast.. The color balance of the original is not 
changed but the brightness of some colors with respect to other colors 
is changed. 

(2) When colored light is used to expose the mask, the mask is 
dense in areas corresponding to the parts of the transparency trans- 
mitting that color, and the mask is transparent in areas corresponding 
to parts of the transparency absorbing the exposing light. In the 
masked transparency and in the duplicate made from it, colors trans- 
mitting the exposing light are made relatively darker and colors 
containing the complement of the exposing light are made relatively* 
brighter. The use of red light for exposing the mask brightens the 
cyans; thus blues and greens appear brighter because they contain 
much cyan. The use of green light to expose the mask would lighten 
the magenta, and the use of yellow (red and green) light would lighten 
the blue by lightening both the cyan and magenta. ^ 

(3) No hue shifts are corrected by a single neutral mask. To cor- 
rect for a hue shift it would be necessary to use the mask for exposing 
only one layer of the duplicating film. This would change the con- 
trast of one component of the neutral scale, and it would then cease to 
be neutral. To restore the scale to a neutral balance would require a 
change in the contrast of one layer of the duplicating film. This can- 
not be done with the present materials and process. 

In order to compensate for all of the optical characteristics of the 
Kodachrome-type dyes that lead to duplicating errors it would be 
necessary to print the duplicate by means of three successive expo- 
sures of red, green, and blue light. Four separate masks would be re- 
quired for the nullification of the four undesirable dye absorptions as 
follows : 

1949 MASKING 149 

(1) Mask of the cyan layer with contrast 0.45 which is capable of 
compensating for the green-light contrast 0f the cyan image. This 
can be made with red light. 

(2) Mask of the cyan layer with contrast 0.45 which is capable of 
compensating for the blue-light contrast of the cyan image. This 
can be made with red light, and in the example can be the same mask 
as (1) because the blue-light contrast is the same as the green-light 

(3) Mask of the magenta layer with contrast of 0.45. This can be 
used to compensate for the blue-light contrast of the magenta image. 
This mask can be made with green light from the transparency with 
the mask (1) in place. Mask (1) will prevent the green-light contrast 
of the cyan dye from influencing the green-light mask. While the 
green-light contrast of the yellow dye will interfere somewhat, the ex- 
tent is insufficient to cause any significant trouble. 

(4) Mask of the yellow layer with contrast 0.15. This can be used 
to compensate for the green-light contrast of the yellow dye. Such a 
mask should be exposed with blue light with masks (2) and (3) in place 
to prevent the blue-light contrast of the cyan and magenta dyes from 
influencing the mask being made with blue light. 

When making the red-light duplicate printing exposure no mask is 
required since only the cyan dye absorbs red, and there are no dye 
additions in the duplicate due to absorptions in other layers. When 
making the green exposure, masks (1) and (4) should be registered 
with the transparency and for the blue exposure masks (2) and (3). 

The net effect of these above masks leaves the neutral scale with 

-YB = 0.60 V G = 0.90 7ft = 1.50 

and the scale is no longer neutral. To make it neutral the processing 
contrast of both the yellow and magenta layers in the duplicate would 
need to be increased to bring them to the same value as the cyan image 
or still further masks would be required to neutralize the neutral scale. 

Full correction by these methods is obviously impractical. 

In a summary of masking for duplicating the transparency onto the 
same material it can be stated that 

(1) A single silver mask provides relative brightness correction, and 
it is a relatively simple technique. 

(2) No hue-shift correction can be accomplished unless the gammas 
of the film layers are changed. This cannot be done in current proc- 
essing procedures. 

150 MILLER February 

(3) Complete correction for unwanted dye absorptions is hypo- 
thetically possible but generally impractical, uneconomical and, in 
fact, usually unnecessary. 


The problem of copying is somewhat more complicated than that of 
duplicating. In making a duplicate the principal objective is to re- 
produce the dye concentrations, and when this is accomplished the 
duplicate will match the original transparency. 

An example of copying is the reproduction of Kodachrome-type 
transparencies on Ektachrome-type film, and the complications which 
arise become apparent on comparison of the dye systems of the Koda- 
chrome and Ektachrome processes. 

The dye system of Ektachrome-type film can be approximately de- 
scribed by Table II. 


Gammas to 
Dyes Blue Green -Red 












A red-light mask in combination with the Kodachrome-type original 
described above nullifies the green- and blue-light contrasts of the 
original cyan. The original cyan, therefore, copies on film of the 
Ektachrome-type as cyan only, no magenta or yellow additions; 
but the copy cyan has different optical characteristics from the original, 
and the two will not match visually. 

Since the magenta and yellow dyes of the original and copy film 
also differ in their optical characteristics, any suitable masking tech- 
nique becomes a matter of multiple masks calculated by much more 
detailed computations than are within the province of this paper. 

Making separation negatives and dye-transfer prints from either 
type original is a third example of copying. In this case it is some- 
what more practical to attempt complete correction of the unwanted 
absorptions: Multiple masks can be used and the contrast of each 
color separation controlled by the development time of the individual 




By using the masks suggested above for the complete correction of 
the unwanted dye absorptions, separation negatives can be made from 
Kodachrome-type transparencies. Each negative is then a record of 
only one of the three images of the transparency. Since the blue-, 
green-, and red-light contrasts of the neutral scale are reduced to 
0.60, 0.90, and 1.00, respectively, the separation negative develop- 
ment is adjusted to bring the contrasts of all negatives to the same 
value. Assuming that a gamma of unity is required, the negatives 
could be developed to gammas of 1.66, 1.11, and 1.0, respectively. 






Fig. 10 Characteristic curves of Ekta- 
color-type cyan-dye scale (broken lines) and 
remaining red-colored cyan coupler (solid 
lines) . 

Neutral-Scale Gammas 

in Masked Kodachrome 


B 0.60 

G 0.90 

R 1.00 

Separation Negative 
Development Gamma 

X 1.66 

X 1.11 

X 1.00 

Contrasts of Separation 



During 1947 the Eastman Kodak Company announced a new 
color-negative sheet film taking material, called Ektacolor. Color 
prints can be made from Ektacolor negatives by the dye-transfer 
process through the use of three separation positive reliefs. Unique 




in the Ektacolor film is the automatic formation, during processing, of 
colored masks within two of the film's layers, to compensate for un- 
wanted absorptions of the cyan and magenta dyes. This represents 
the most recent advance in the science of masking. While Ektacolor, 
as announced, will be supplied only as a portrait film, the possibility 
of the basic principle's being extended to other color materials merits 
its description here. 

Ektacolor is a three-layer film with the emulsions sensitive to blue, 
green, and red. Incorporated in the respective layers are yellow- , 
magenta- , and cyan-forming couplers. In the first film-processing 





Fig. 11 Characteristic curves of Ektacolor- 
type magenta-dye scale (broken lines) and re- 
maining yellow-colored magenta coupler (solid 

step three negative silver images are simultaneously developed in 
the layers. The developer is oxidized in proportion to the silver de- \ 
veloped, and the developer reaction products combine with the ; 
couplers to produce negative dye images. 

After the silver and unused silver halide are removed from the film, ; 
only the three negative dye images and the unused coupler remain, j 
The coupler that produces cyan dye is reddish in color and the I 
coupler that produces magenta is colored yellow. Since the original j 
color of the couplers is destroyed as the new image dyes are formed, j 
the remaining couplers comprise positive colored images. These are j 
the masks. 




The dye images formed from colored couplers in the Ektacolor-type 
film exhibit optical characteristics similar to those formed from color- 
less couplers in Ektachrome-type films. For photographic reasons, 
however, the over-all contrast of the Ektacolor-type is kept lower 
than the Ektachrome-type. 

If the colorless coupler-dye gammas were adjusted, for example, as 
follows, a neutral scale with a gamma of 1.0 would result. The table, 
therefore, indicates the magnitudes of the corrections which must be 
accomplished by the colored coupler masks. 








Gammas to 







/ _____ ""VJ5 


Fig. 12 Characteristic curves of Ektacolor- 
type yellow-dye scale. 

The reddish-colored positive image formed by the unused cyan 
coupler cancels the negative blue- and green-light contrast of the cyan 
dye, and the yellow-colored positive image formed by the unused 
magenta coupler cancels the negative blue-light contrast of the 
magenta dye. At present there are no entirely suitable magenta- 




colored yellow couplers so the green-light contrast of the yellow dye is 
not masked. 

Cancellation of the unwanted absorption values in the cyan and 
magenta layers will leave the total red-, green-, and blue-light absorp- 
tions unbalanced. The manufacturing and film processing are, there- 
fore,, adjusted to give a balanced neutral scale when the unwanted 
cyan and magenta absorptions are masked. With the increased con- 
trast in two of the layers, the dye absorptions in the present example 
are, therefore, approximately: 







Gammas to 






Fig. 13 Characteristic curves of Ektacolor- 
type neutral scale measured by blue, green, 
and red light. 

In order to be completely effective as a mask for the cyan layer, the 
remaining reddish coupler must have blue-light contrast of 0.10 and 
green-light contrast of 0.20 with maximum densities to these two 
colors equal to the maximum densities of the respective absorptions of 
the cyan dye. These characteristics are provided and controlled 
during the making of the couplers (Fig. 10). 

1949 MASKING 155 

To be effective as a mask of the magenta layer the remaining yellow 
coupler must have a gamma of 0.30 to blue light and a maximum 
density to blue equal to the maximum density of the magenta dye to 
blue light (Fig. 11). 

There is no correction in the yellow layer (Fig. 12). 

With the corrections afforded through the use of colored couplers, 
the color negative closely resembles the ideal dye image character- 
istics (Fig. 13) . Exceptions are the very slightly low magenta contrast 
and the equally slight green absorption of the yellow dye. From such 
a corrected negative separation, positives and color prints can be made 
substantially free of errors due to the unwanted dye absorptions. 


Fortunately it is easier to make a mask and use it in duplicating and 
copying (except for motion pictures) than it is to understand the exact 
reasons for its use. The manufacturers of color films issue instruc- 
tions for masking in some detail, and these generally provide the best 
possible practical method for obtaining adequate color reproduction. 
It is important, though, for the color technician to understand the 
processes with which he is working, and to know their possibilities and 
limitations; only then can he make the best use of the instructions 
and materials which he has at his command. 


The presentation of this paper at the Convention was illustrated 
with 73 color slides and 14 large color transparencies. The author 
acknowledges, with thanks, the contributions of those from various 
departments of the Eastman Kodak Company who provided data and 
laboratory assistance to make the paper and illustrations possible. 


The following three papers were presented 
at the 1 63rd Semiannual Convention of the 
Society of Motion Picture Engineers at 
Hollywood in May, 1948, at a joint meeting 
with the Inter-Society Color Council. Two 
other papers presented at the same sessions 
were most instructive, but consisted of 
physical demonstrations and color illustra- 
tions to such an extent that reproduction in 
the JOURNAL is impractical.* 

To many of our readers the Inter-Society 
Color Council, which so generously prepared 
the program for the joint meeting, needs no 
introduction. However, for the benefit of 
others not acquainted with the ISCC, it may 
be noted that it is not just another technical 
society. Its members are mainly technical 
societies or associations having color prob- 
lems, such as the American Pharmaceutical 
Association, Illuminating Engineering So- 
ciety, and eleven others. The SMPE is a 
member, being represented by a committee 
of which Ralph Evans is chairman. 

The purpose of the ISCC is "to stimulate 
and co-ordinate the work being done by 
various societies . . . leading to standardiza- 
tion, description, and specification of color, 
and to promote the practical application of 
these results ."..'." In accordance with this 
purpose, the SMPE will make the three fol- 
lowing papers available. 


* "Color Phenomena," by Isay Balinkin, and "Seeing 
Light and Color," by Ralph M. Evans. 

Spectral Characteristics of 
Light Sources* 






Summary New light sources are being developed regularly. No longer 
can industry depend upon carbon-arc lamps and incandescent lamps alone. 
A brief description of all the important light sources will be made with special 
emphasis on their spectral characteristics and their effect on colored objects. 


EFFICIENCY AND SIZE of light sources have an important bearing on 
the selection of lamps for motion picture photography, yet more 
and more, where color effects are desired, their color and spectral 
characteristics become important. Natural daylight, the carbon-arc, 
and incandescent tungsten-filament lamps provide most of the illumi- 
nation today for the motion picture industry. While they remain 
the most important sources, there have been other developments 
during the past ten years that for one purpose or another have a 
special bearing or interest for the motion picture engineer. It is our 
purpose in this report to assemble as much typical information as is 
available on the spectral characteristics of both the standard sources 
and on the more newly developed fluorescent, mercury-cadmium, and 
concentrated zirconium-arc sources. 

First, however, attention should be called to the fact that the 
shorthand method of specifying the color of illuminants in terms of 
color temperature is a practice that often obscures the differences 
between color and spectral composition. For that reason it seems 
important to indicate that the Planckian locus represents the locus of 
chromaticities of a black-body radiator at various temperatures, 
Fig. 1, and that it is only necessary according to American practice 

* Presented May 20, 1948, at the SMPE Convention in Santa Monica. 





in order to apply the term* color temperature to a source, that it 
match the color of a black body at a given temperature. It is not 
necessary that it should match the black body in spectral distribution, 
Fig. 2. 

The specification of the color of sources in terms of color tempera- 
ture without an understanding of the limited meaning of the term has 
caused much confusion in color thinking as it concerns the illuminant. 

Spectrum locus 




Fig. 1 The Planckian locus on a chromaticity diagram 
consists of points that represent chromaticities of a black- 
body radiator at various temperatures. 

It should therefore be pointed out that the Optical Society of America, 
the American Psychological Association, and the Illuminating Engi- 
neering Society definitions 1 of color temperature refer to chromaticity 
only.* The British, on the other hand, in a definition 1 proposed by 

* A note attached to the IES definition states that the term color temperature is 
"usually assignable only for sources which have a spectral distribution of energy 
not greatly different from that of a black body." In practice the IES does not 
hold to this restriction. 




the Colour Group of the Physical Society, require that a source be of 
substantially the same spectral distribution in the visible region as a 
full radiator of the same color. Thus, the Americans and British 
differ fundamentally in their concept of the term. 

The meaning and use of the term color temperature as a short-cut 
method for describing the chromaticity of an illuminant should be 



Fig. 2 Spectral-energy distribution in the visible 
region of the spectrum for a black body radiating at 
various temperatures. 

clearly kept in mind for what it is. For any real understanding of 
color processes, whether visual or photographic, it is necessary to take 
into consideration the more exacting specification of spectral distribu- 
tion. Thus, while illuminants in this report are often referred to in 
terms of the color-temperature scale, it should be remembered that it 
is not their color but only their spectral characteristics that will tell 
whether they are suitable for use with a given film, or to produce a 
specified result. 


Another point that should be emphasized if we are to continue to use 
the term color temperature is that reciprocal temperature provides a 
better scale for expressing chromaticity differences than does tempera- 
ture itself. A chromaticity difference of 100 degrees Kelvin at 3000 
degrees Kelvin is a very important color difference but at 6500 degrees 
Kelvin it is hardly significant. In 1933 I. G. Priest 2 proposed that 
reciprocal temperature be adopted as the conventional parameter for 
specifying the chromaticity of incandescent illuminants and various 


Fig. 3 ICI (x, y) diagram showing a portion of the Planckian locus in the 
range 2848 degrees Kelvin (A illuminant), through B and C illuminants to 
above 10,000 degrees Kelvin. Intervals shown are 20 mireds (/xrds) from 400 
to 100, with equivalent color temperatures indicated. 

phases of daylight and that the microreciprocal degree absolute centi- 
grade (abbreviated "mired," or "jurd") be adopted as the most con- 
venient unit. There are several reasons for this proposal, all valid, 
as anyone will find who works with color difference specifications that 
relate to illuminants. 3 On Fig. 3 the intervals from the color tempera- 
ture of illuminants from 3000 to 10,000 degrees Kelvin are 20 mireds; 
compare this with the widely varying size of intervals when color 
temperature is used, as in the 1000-degree-Kelvin intervals on Fig. 1. 
This would be quite as evident on a uniform chromaticity diagram. 


Regarding spectral distribution of light sources there are a number 
of reports that may be referred to for information of general interest 
in the motion picture field. The 1943 report of Linderman, Handley, 
and Rodgers 4 is well illustrated to show various lighting techniques. 
It discusses lamp requirements for color quality of light and for 
spectral distribution of film and includes a detailed listing and de- 
scription of 22 types of carbon-arc and incandescent lamps used in 
motion picture studios. The SMPE Studio Lighting Committee 
Report of 1945 5 deals with operation and maintenance of carbon-arc 
and incandescent lamps, the section on incandescent lighting includ- 
ing a discussion of color temperature as one of the variables. A great 
deal of quantitative data for computational use in studying the color 
of illuminants and their effect on object colors is contained in the 
report on Quantitative Data and Methods for Colorimetry published 
in 1944 by the Committee on Colorimetry of the Optical Society of 
America. 6 Both this and the Linderman report contain a number of 
good references. Tables containing spectral radiancy and spectral- 
distribution data for computing ICI colorimetric values for black- 
body radiators at temperatures 2800 to 3800 degrees Kelvin for every 
100 degrees Kelvin have been published by Adams and Forsythe, 7 but 
perhaps the most extensive and available tables of colorimetric com- 
putational data for Planckian radiators and a variety of actual 
sources, for use either by the weighted or selected ordinate method of 
calculation, are contained in the report of the OSA Committee on 
Colorimetry. 6 * 

References devoted to specific illuminants are included in the 
appropriate section of the discussion that follows. 

Since the authors have prepared this paper as part of a technical 
session on color arranged by the Inter-Society Color Council, there are 
other references that should be included at this point. Such refer- 
ences may not belong in a paper strictly devoted to a discussion of 
spectral characteristics of light sources, but the questions they raise 

* Since the Santa Monica meeting at which this paper was given, a new book by 
Forsythe and Adams 8 has come to the authors' attention. While one might infer 
from the title that the book refers solely to fluorescent and other gaseous-discharge 
lamps, attention should be called to the fact that it includes a ch'apter on arcs of 
various types, one on light sources of short duration, one on delayed phosphores- 
cence, and one on fluorescence and television. Any book on lamps or lighting by 
these authors is welcome, and to find one that brings the subject up to date in 1948 
is particularly welcome, and should prove useful to any motion picture engineer 
interested in the spectral and other characteristics of current light sources. 


must be considered before a motion picture or any other engineer can 
understand and control the direction and size of color change with 
change of illumination, and his attention should therefore be called 
to them. These questions involve studies of surround, adaptation, 
constancy studies that fall more in the province of the psychologist 
than the physicist. Yet such questions must be studied in the future 
with quite as much care and in as much scientific detail as has been 
devoted in the past two decades to problems of color measurement 
and specification. 

There is, for example, a book by Katz, 9 written in 1911 and trans- 
lated from the German in 1935, that discusses various factors involved 
in studies of color constancy. More recently we have in this country 
the work of Helson, Judd, and Evans. The work of Helson and his 
students and associates at Bryn Mawr has resulted in a series of five 
papers already published, 10 " 14 one paper in press 16 and two Ph.D. 
dissertations to be published. The last two papers concern change in 
hue and saturation of aperture colors as a function of the composition 
and luminance of the surrounding field (R. V. Higbee) and the effect 
of general and local shadows on hue, lightness, and saturation (J. deB. 
Brugger) . A study to find a formula for predicting the changes found 
in the Helson experiments has been made by Judd, 16 who expects to 
continue the work until a formula is found to fit all the requirements. 
Still more recently we have the new Evans book 17 which calls atten- 
tion to many general illumination problems, just as he has been calling 
the attention of the SMPE and other groups to them during the past 
ten years in a series of well-illustrated lectures. Color needs to be 
seen if its effects are to be intelligently studied, and Ralph Evans' 
illustrated lectures have enabled more people to gain a picture of some 
of the many color problems not yet understood, certainly not yet 
answered, than has been possible by any other method of presentation. 
His book, which is well illustrated in color, should reach a wide audi- 
ence. It is to be hoped that it may awaken many a young student of 
color to the possibilities that lie ahead in this field of color research. 

While these few publications do not provide an answer to the entire 
problem, they do call attention to phases of it that are often omitted 
from consideration during technical color training. It is only when 
one comes up against practical problems that he finds there are these 
other matters that at present must be handled by experience or on a 
basis of trial and error. Actually, if results already published by 
Helson, Judd, and Evans, and in Holland by Bouma and Kruithof , 18 - 19 




were more widely understood, a great many practical answers could be 
worked out today. These references are therefore included so that 
the motion picture group may have its attention particularly called 
to this type of study. They relate to color and illumination problems 
encountered every day in the motion picture industry. 


Scientists have been trying for years to equal the color of natural 
illumination witji reasonable efficiency. The spectral-energy dis- 
tribution curves of daylight and sunlight are enough like those of a 
theoretical black-body radiator so that black-body-distribution curves 
are often used as standards for comparison. For a discussion of 





200 nr 


Wave length (A) in 

Fig. 4 Reference curves for energy distribution in the visible region for 
three types of standards used in studies of natural day lighting: (A) curves 
of natural sunlight and daylight; (B) ICI standard illuminants for colori- 
metric work; (C) black-body distributions in the 6500- to 8000-degree-Kelvin 

spectral emittances of complete radiators, and a list of references to 
tables and charts specifying relative values for various temperatures, 
see 20 section 18. There are certain exceptions to this resemblance, 
however, for absorption of energy in the earth's atmosphere, due 
chiefly to the presence of moisture, causes some differences between 
these curves as shown in measurements by Abbot, 21 and in proposals 
of Moon 22 for standard curves for engineering use. As standards for 
colorimetric work the International Commission on Illumination in 
1931 adopted three illuminants, known as ICI Illuminants 23 ' 24 J., B, 
and C. Illuminant A represents. a tungsten-filament lamp that oper- 
ates at a color temperature of 2848 degrees Kelvin with an energy 
distribution very like that of a black body ; the other two are intended 
to represent two phases of natural daylight, B that of noon sunlight, 




C that of average daylight. They are produced by using specified 
liquid filters with lamp A. For comparison these three types of 
standards used in studies of natural daylighting are shown in Fig. 4 
those for natural daylight, for ICI standards, and for Planckian or 
black-body distributions in a range of color temperatures 6500 to 
8000 degrees Kelvin. 

It is difficult and time-consuming to make energy measurements of 
natural daylight; therefore, few data are available. Yet the color of 



600 600 700 400 600 60.0 



Fig. 5 Curves approximating spectral-energy 
distribution of daylight; On the left, as calculated by 
Gibson for a range of Abbot daylight to a limit blue 
sky; on the right, as measured by Taylor in 1939. 

daylight may change through a wide gamut from sunrise to sunset, 
and from sunlight to blue sky, or to daylight provided by cloudy skies. 
K. S. Gibson of the National Bureau of Standards some years ago 
suggested a formula 25 by which the spectral energy of sun-outside-the- 
atmosphere (he used Abbot data) may be combined with the Rayleigh 
scattering equation to provide an approximate spectral-energy dis- 
tribution for any given color temperature of skylight. Fig. 5 shows 
curves of this sort for color temperature of Abbot's daylight-outside- 
the-atmosphere (about 6000 degrees Kelvin), up to a limit blue sky 




(at oc ) . Paired with this diagram is one showing measurements made 
by Taylor and Kerr 26 in 1939 of daylight ranging from 4975 up to 
60,000 degrees Kelvin. 

Measured color temperatures of natural daylight are reported up to 
60,000 degrees Kelvin. On the other hand late-afternoon sunlight 
may measure as low as 2000 degrees Kelvin, 27 actually lower than the 
color temperature of most incandescent lamps which for ordinary use 
usually range from 2800 to 3000 degrees Kelvin (depending on size), 








Fig. 6 Spectral-energy-distribution curves: (A) For tungsten filaments 
operating at equal voltage but different temperatures (from Linderman, 
Handley, Rodgers report; 4 () For Planck's formula, adjusted at 560 to 
the tungsten-filament values shown for (A). (The actual temperature of 
the tungsten filament is lower than that of a true black body when they are 
a color match.) 

and for photographic work up to as much as 3400 degrees Kelvin for 
lamps of fairly short Me. 

Local atmospheric conditions may cause the color and intensity of 
natural light to vary considerably especially in cities where smoke and 
fumes may serve to filter the daylight, and result sometimes in major 
deviations from expected energy distributions of natural daylight. 

Most manufacturers of color film have developed their emulsions so 
that they will be exposed either in a high color temperature of incan- 
descent tungsten (3200 degrees Kelvin) or in an overcast daylight of 
about 6500 degrees Kelvin. For daylight film the best results are 
obtained when the daylight most closely resembles the color and 




spectral distribution for which the film was designed.* Practically 
no combination of cloud effects and daylight can produce the dis- 
astrous resultant color effects on color film that are produced by sharp, 
bright, spectral lines, such as those of the ordinary mercury lamp. 


The oldest sources, yet those most regularly used by motion picture 
engineers, are incandescent tungsten filaments and carbon arcs. 
Great progress has been made in both. 

In the case of incandescent tungsten filaments, whose character- 
istics have been thoroughly described by Forsythe and Adams, 28 the 
early low-wattage lamps are now replaced by 5000- and 10,000-watt 
lamps that have increased both in color temperature and in lumen-per- 
watt efficiencies. 


K . 5KW T-64(G-64> LAMP 







Fig. 7 Color-temperature characteristics of 5-kilowatt studio 
lamps during life applicable to lamps that are cleaned at least every 
20 hours of burning by swirling tungsten cleaning powder about in 
the bulb. (From SMPE Studio Lighting Committee Report. 6 ) 

This source is admirably suited for motion picture work because of 
dependability and low labor maintenance. Especially attractive is 
the close approximation of the spectral-energy distribution of the in- 
candescent lamps to that of a black body at equivalent color tempera- 
ture. Spectral-energy-distribution curves for tungsten filaments 
operating at equal wattage but different temperatures are shown in 
Fig. 6A. Paired with the distribution curves in Fig. 6A is a diagram, 
Fig.-6B, showing a series of distributions according to Planck's form- 
ula, adjusted at X 560 to match the corresponding data for the 

* Spectral sensitivities for various types of color film must be known before 
accurate colorimetric calculations can be made to select the illuminant of most 
suitable spectral distribution for use with a particular color film. 




tungsten-filament lamps. As may be seen, the relative distributions 
for incandescent tungsten filaments and for black-body radiators at 
equivalent color temperature are similar. The actual temperature 
of the tungsten filament is lower than that of a true black body 
when they are a color match, about 100 degrees at 3000 degrees of 
color temperature. 

The change in color temperature of lamps operated at constant 
voltage has been discussed by Judd 29 for certain lamps. The change 



^/ATTS / 

















I 60 

, ^ 































Fig. 8 Color temperature and c 
filled tungsten lamps when operated 
SMPE Studio Lighting Committee ! 

8 Color temperature and other changes characteristic of gas- 
" at other than rated voltage. (From 
Report. 6 ) 

in 5-kilowatt studio lamps, applicable to lamps cleaned every 20 hours 
of burning by swirling tungsten cleaning powder in the bulb, was 
reported by the SMPE Studio Lighting Committee in 1945. 5 Fig. 7 
is taken from their report. Fig. 8, also taken from their report, 
indicates the color-temperature change that may be expected of 
lamps operated at other than rated voltage. 

Lamps for black-and-white photography vary in color temperature 
from 2900 degrees Kelvin for 75-watt lamps to 3350 degrees Kelvin 
for 10,000-watt lamps and are used chiefly for other characteristics 




than color temperature ; for color photography the color temperatures 
are important, and usually fall into two classes, either 3200 degrees 
Kelvin intended for use with color films such as Eastman's Type B 
or Ansco's Tungsten Type or 3200-degree-Kelvin film, or 3350-degree- 
Kelvin lamps, often called "CP" lamps. 



I 60 



I 40 

Limit Filters 


500 600 



Fig. 9 Spectral transmissions of Macbeth Whiter- 
lite filters used with high-temperature incandes- 
cent lamps to provide a daylight distribution suit- 
able for use with Technicolor film. The two curves 
represent tolerance limits. 

Films such as Technicolor are made for daylight only. For such 
films the incandescent lamp can still be used, usually the 3350-degree- 
Kelvin lamp, in that case with a color-correcting filter suitable for 
maintaining a spectral distribution of illumination that is required 
for use with such film.* A film such as Technicolor requires some- 

* If films were sensitive to the same wavelength distribution as the average human 
eye, then daylight itself, or as good an artificial daylight as possible, should be used 




what more transmission in the red than would be supplied by daylight 
filters designed for visual work. This excess in the red, while keeping 
the rest of the curve close to that of daylight, is supplied by Macbeth 
Whiteiiite filters. Spectral transmissions of these filters are shown 
in Fig. 9. The two curves are those for tolerance filters supplied in 



+ 3350K Source 


500 600 



Fig. 10 Relative spectral-energy-distribution curves 
of 3350-degree-Kelyin tungsten-plus-Whiterlite limit 
filters. These distributions provide limit color tem- 
peratures of 6300 to 7100 degrees Kelvin. 

for daylight film. The closest filtered approximation to daylight employs tungsten- 
filament lamps with Macbeth Daylite filters (Corning 5900) . 3 - 30 ~ 32 The thickness 
of a given filter controls the color temperature through a range from about 5000 de- 
grees Kelvin to 8500 degrees Kelvin when used with lamps of required wattage. 
Such filters as these are supplied for visual tasks where carefully standardized arti- 
ficial daylight is a requirement. 33 ' 34 Films, however, are not sensitive to the same 
distribution as the eye, and may require adjustments in transmission for certain por- 
tions of the spectrum in order to compensate for this difference. We note this dif- 
f ence in order that there may be no confusion between the Daylite filters used for 
visual work and the Whiterlite filters supplied for use in color photography. 




accordance with specification, graded according to transmission. 
One of the filters, 145 per cent (American Association of Railroads), 
produces a calculated color temperature of 7100 degrees Kelvin when 
used with a 3350-degree-Kelvin lamp. The second filter, 165 per 
cent (American Association of Railroads), produces 6300 degrees 


180 - 

160 - 


O 120 






Type 170 without filter-5820K 




Fig. 11 Energy distribution of light from carbon-arc lamps used in studio 
lighting. These curves represent the light emerging from the optical system: 
Type 40 Duarc at 4710 degrees Kelvin; Type 170 without filter at 5820 degrees 
Kelvin, with filter at 5070 degrees Kelvin. 

Kelvin with the same lamp. Relative-energy-distribution curves for 
a lamp-and-filter combination are shown in Fig. 10. 

This combination of lamp and filter has the low labor maintenance 
factor of incandescent. However, the efficiency in lumens per watt 
of the incandescent lamp is reduced nearly two thirds by filter absorp- 
tion, which is a subtractive method of producing color. The resulting 
energy distribution is, however, well suited for Technicolor. Filter 


holders designed for use with spot-type illuminators using incan- 
descent sources have been available for some time. A new develop- 
ment is a broad light fitted with filters. 


Because the high-temperature carbon arc provides a great quantity 
of illumination from a single source and also provides a sufficiently 
close approximation to the spectral-energy distribution of sunlight, it 
is the most generally used source of illumination in the motion picture 
studios for color work. It has the advantage of being a radiating 
source which, by use of various chemical combinations, can supply 
good spectral characteristics that resemble the continuous curve of a 
black-body radiator. Color temperatures varying from 3500 to 3900 
degrees Kelvin for low-intensity carbon arcs, and from 4500 to 6500 
degrees Kelvin, high-intensity carbon arcs can be produced. 

Spectral-energy-distribution curves of several types of carbon arcs 
are illustrated in the Linderman, Handley, and Rodgers report. 4 
These are characteristic of the arc sources as directly viewed without 
the modification which occurs as a result of use with an optical system. 
In Figs. 11 and 12 the curves shown are characteristic of the radiation 
in the light beams as they emerge from the optical system. Fig. 11 is 
a modification of Fig. 8 of a paper by Bowditch, Null, and Zavesky. 86 
This shows the energy distribution of the most popular units employed 
in studio lighting, the Type 40 JDuarc at 4710 degrees Kelvin which 
seems quite satisfactory either alone or mixed with sunlight for pro- 
viding a proper balance for use with Technicolor, and the Type 170 
lamp at 5820 degrees Kelvin, which when used with a filter gives 
5070 degrees Kelvin, and produces a suitable spectral distribution.* 

Fig. 12 shows the spectral-energy distribution of the light on the 
theater projection screen with two high- and one low- intensity pro- 
jection arcs; low-intensity carbon arc at 3870 degrees Kelvin, 

* The fact that Technicolor film is adequately served by tungsten-plus- Whiterlite 
filters within color temperature limits of 6300 to 7100 degrees, and by carbon arc 
reduced from 5820 to 5070 degrees Kelvin by use of a filter, is a clear demonstration 
of the fact, discussed in the introduction to this paper, that color temperature may 
often prove to be a very unsatisfactory measure of the color characteristics of light 
sources, particularly when light sources depart substantially from the spectral 
distribution of a full radiator of the same color. Experience shows that spectral 
distributions of either the tungsten-plus- Whiterlite filter or indicated filtered car- 
bon arc provide satisfactory results with Technicolor, yet the color temperature of 
one averages 6700 degrees Kelvin, the other just over 5000 degrees Kelvin. 


"Suprex" carbon arc at 5380 degrees Kelvin, and a 13.6-mm High- 
Intensity Projector carbon arc at 5600 degrees Kelvin. 

When filters are used with carbon arcs, it has usually been supposed 
that they are necessary to take care of the peak energy at about 390 
millimicrons considered characteristic of this source. However, as 
Bowditch points out, the 390-millimicron peak is apparently a char- 


180 - 









13.6mm H. I.-5600K 


^ Low inteniit 

Low intnity-3870 K 

300 400 500 600 700 


Fig. 12 Energy distribution of light from carbon-arc lamps used for pro- 
jection. One low- and two high-intensity lamps are illustrated, low intensity 
at 3870 degrees Kelvin, Suprex at 5380 degrees Kelvin, and a 13.6-mm high- 
intensity at 5600 degrees Kelvin. 

acteristic of the diffuse arc flame which is not effectively projected by 
the optical system. The filters are used to reduce the intensity in the 
far -blue and near ultraviolet so as to provide a better balance with 
color processes designed for daylight. A gelatin-straw filter may be 
used, also a new glass coating which can be applied even to Fresnel 
lenses to supply color correction without the fading which character- 
izes the use of gelatin. 


All of its advantages, which include high efficiency at temperatures 
much above those of tungsten, make the carbon arc a most important 
source in the motion picture industry. Its high brightness, plus its 
color and efficiency, makes it the standard source for theater pro- 

Bowditch and his associates, who have discussed the spectral char- 
acteristics of carbon arcs before this Society at various times, are 
authorities in this field and their reports may be referred to for further 
information. 35 " 39 


The oldest of the new and radical deviations from standard light 
sources such as have been described is the flourescent lamp. It is ten 
years 8 - 40 since this new source was announced. It offered, and still 
offers, exciting adventures in color through the use of fluorescent pow- 
ders. Between 80 and 90 per cent of the total lumen output of a 
fluorescent lamp is derived from fluorescence; the balance is mercury 
light transmitted through the phosphors coated on the inside of the 
glass tube. To date, fluorescent lamps have been made available in a 
variety of colors blue, green, pink, red, gold but since our interest 
is primarily in sources that provide colors reasonably close to those of 
black-body radiators, this discussion will be confined to those fluores- 
cent lamps that are called white or daylight in color and are usually 
assigned a color temperature by the manufacturers. 

In Fig. 13 relative spectral-energy-distribution curves of three 
standard fluorescent lamps are shown: 3500 degrees Kelvin, 4500 
degrees Kelvin white, and 6500 degrees Kelvin "daylight." If the 
spectral-energy distributions of these lamps are compared with 
natural daylight, or of black-body radiators at similar color tempera- 
tures, it will be found that these curves fall short in three respects. 
First, they are low in energy in the red end of the spectrum because, 
until recently, no phosphor has been available which could be acti- 
vated sufficiently in the red region of the spectrum between 650 and 700 
millimicrons. Second, they are low in the blue end because no blue 
phosphor of sufficient strength in the 400- to 420-millimicron region 
is available. Third, about 10 to 15 per cent of the radiation of these 
lamps is in sharp, bright mercury lines which distort color, color film 
in particular. 

Recently, because of the very great importance to the photographic 
industry of correcting these deficiencies, particularly in the red, a new 




phosphor has been developed for use in the red end. This has been 
studied by Froelich and described in a paper by Buck and Froelich. 41 
It is a double-activated calcium phosphor with a two-peak emission, 
one at 650 millimicrons with 50 per cent intensity at 700 millimicrons, 
another at 360 millimicrons. The best initial efficiency obtained with 
this phosphor in a 40-watt lamp was about 12.5 lumens per watt, 





( 35000 K) 


500 600 



Fig. 13 Spectral-energy-distribution curves for 
three standard fluorescent lamps : 3500 degrees Kelvin 
white, 4500 degrees Kelvin white, and daylight 6500 
degrees Kelvin. 

which is low as compared with the 65 to 70 lumens per watt with the 
phosphor for the red now used in the white lamps. Unfortunately, 
only one manufacturer has announced a fluorescent lamp employing 
this new phosphor in lamps for general use, although others provide 
it for restricted use, as in photocolor fluorescent lamp. This is due to 
the fact that while this red-corrected fluorescent lamp provides a much 




closer approximation to tjie spectral-energy requirements for film and 
visual applications, the addition of this red phosphor lowers the over- 
all efficiency of the fluorescent tube enough so that manufacturers at 
the present time are convinced that maintaining present efficiencies is 
more important than the color improvement which they consider to be 
minor. There is no reason to believe that a new blue phosphor can- 
not be developed. 

It may, however, take some tune before phosphors of sufficient 
sensitivity are developed which will permit a heavy enough coating on 
the tube to absorb 100 per cent of the visible mercury radiation and 
still permit the ultraviolet radiation to excite the phosphors that lie 
on the glass surface, and thus produce sufficient visible radiation. 







Fig. 14 Spectral-energy-distribution curve of 5000-watt "compact- 
source" mercury lamp. (Adapted from curve supplied by the British- 
Thomson-Houston Company.) 

The size and large numbers of lamps needed to provide sufficient 
illumination for motion picture work, and the difficulty of applying 
optical controls, would continue to restrict their use for many pur- 
poses even if the spectral characteristics of these lamps were much 


For many years, lamp manufacturers have been developing higher- 
pressure mercury lamps. Research has, quite properly, been directed 
toward this end, since, especially for studio lighting, a high-brightness 
source is required and mercury 'lamps provide this feature. 42 " 44 Be- 
cause of its spectral distribution, however, a straight mercury source 
has been poor even for black-and-white photography. The spectral- 




energy distribution of a 5000-watt compact-source mercury lamp is 
shown in Fig. 14. The objectionable features include the bright mer- 
cury lines, and the great gaps in the visible region where the virtual 
lack of energy makes certain sharp cutoff colors reproduce as black or 
gray, while colors having a dominant wavelength on one of the mer- 
cury lines reproduce as overly bright. 

As reported by Carlson 45 at a recent SMPE meeting, during the war 
considerable work was done in England by the British-Thomson- 
Houston Company on the combination of mercury and cadmium. As 
may be noted in Fig. 15, almost double the number of lines or bands 
are produced by this combination as for mercury alone. These lines 






ig. 15 Spectral-energy distribution of mercury-cadmium lamp. (Adap- 
from curve supplied by the British-Thomson-Houston Company.) 

are well distributed over the visible region. However, these lines or 
bands still leave large gaps where little radiation is present, and thus 
still produce unnatural effects on both colored objects and color film. 

Although this source may prove good for black and white providing 
a reasonable cost per unit can be established, for color photography 
there seems to be a conflict of opinion. Theoretically, unless new 
metals can be found which will create new bands of approximately 
equal brightness over the whole visible spectrum so as to simulate a 
continuous black-body radiator, satisfactory results are not be be ex- 

While visiting England last year, the senior author of this report 
was told of questionable color results obtained by users who tried this 
source for color photography. 



One of the latest lamp developments is the concentrated zirconium- 
arc lamp developed by the Western Union Telegraph Company and 
reported two years ago to the Optical Society of America and to the 
Society of Motion Picture Engineers by Buckingham and Dei- 
bert. 46 ' 47 




500 600 


Fig. 16 Spectral-energy distribution of concentrated 
zirconium arc lamp. (Adapted from curves shown by 
Buckingham and Deibert. 46 - 47 ) 

Because the concentrated arc is a hot body, it is actually a continu- 
ous radiator over the entire visible spectrum. As may be noted from 
Fig. 16, its relative-energy distribution is smooth throughout most of 
the visible spectrum, providing a color temperature of about 3200 
degrees Kelvin. Below 400 millimicrons and above 695 millimicrons 
there are several bands of energy caused by gases, but these are not 
evident in the visible portion of the spectrum, and should not inter- 
fere in color photography except to film sensitive in those regions. 


While this concentrated arc is new and not yet fully developed so as 
to be placed in low-cost manufacturing, it has been used already in 
optical equipment because of the fine definition provided by the small 
point-type source of the lower-wattage lamps. 

There is no reason why this 3200-degree-Kelvin source cannot be 
raised in color temperature by use of filters, just as incandescent is 
used with filters for Technicolor. 

An important feature of the concentrated arc, both spectrally and 
photographically, is that the change of color temperature caused by 
voltage fluctuations is very small, and shows no greater change during 
life. Over a very wide voltage fluctuation (60 to 140 volts) for the 
300-watt lamp, the color-temperature range is no more than 200 
degrees Kelvin, and for normal voltage fluctuation of 5 or 10 volts, is 
very small indeed. 

Although originally produced in sizes ranging from 2 to 100 watts, 
further developments during the past two years have been directed to 
the production of lamps larger than 100 watts. A number of major 
problems have had to be solved in their design and construction. A 
300-watt lamp is now available commercially, a 1000-watt concen- 
trated arc lamp is developed but not yet in commercial production, 
and work is proceeding on the design of a 5-kilowatt lamp. 

The 300-watt lamp is provided with two anodes of molybdenum 
rod, */4 inch in diameter by 4 inches long. In normal operation the 
brightness is around 50 candles per square millimeter, the lamp 
showing an efficiency of 1.8 candles per watt, about twice as efficient 
as the 100-watt concentrated arc lamp. One satisfactory design of 
the 1000-watt size shows the lamp contained in a 6-inch diameter 
bulb, with the cathode spot about 2 3 /4 inches from the bulb wall. 
As reported by W. D. Buckingham,* lamps of this type have had lives 
of several thousand hours in the laboratory. A few lamps made with 
a 4-inch diameter bulb appear also to have a reasonably long life. 
The candle power of these lamps, when operating at their normal 
current (50 amperes), is about 1600, the brightness centers around a 
value of 50 candles per square millimeter, the area of the spot is nor- 
mally about 35 square millimeters, and their efficiency is about 1.3 . 
candles per watt which is somewhat less than that of the 300-watt 
lamp. The 1000-watt lamp is expected to find application in the 

* Reported by W. D. Buckingham at a conference on High-Intensity Light 
Sources, Northwestern University, June, 1948. 




field of 16-mm motion picture projection, for the image of a rectangu- 
lar cathode spot can be placed directly at the film gate of the pro- 
jector with little loss of light. The design being followed in the 
development of 5-kilowatt lamps is similar to that of the 1000-watt 
size but the electrodes are larger and the anode made with four radiat- 
ing fins. The luminous spot will probably be slightly over one-half 
inch in diameter. 

Progress is also reported on the development of brighter lamps. 
Lamps twice as bright, but considerably less efficient, have been pro- 
duced by use of hafnium oxide as a cathode-filling material. The 



| 60 









500 600 




500 600 



Fig. 17 Samples represented by the curves in A are a color match under 
ICI illuminant A. Those represented in are a color match under ICI illumi- 
nant C. The colors of these pairs will differ greatly when one illuminant is 
substituted for another. 

spectra produced by the hafnium and zirconium lamps are quite 
similar. The real advantage in the hafnium-type lamp, by compari- 
son of the 1000-watt size, is that its brightness varies between 90 and 
110 candles per square millimeter, while the brightness of a zirconium 
lamp usually lies between 45 and 55 candles per square millimeter. 
The efficiency of the 1000-watt hafnium lamp is 0.9 compared with 
1.3 for the zirconium lamp. The cost and shortage of hafnium oxide 
is delaying commercial production of these newer type lamps, al- 
though experimental lamps are available. 



In place of demonstrations of the effect of illumination that were 
a part of this paper as presented at the Santa Monica meeting, spec- 
trophotometric curves of two pairs of samples are shown in Fig. 17. 
The importance of illumination in effecting color changes is indicated 
by the fact that the samples represented by two curves in Fig. 17A 
are a color match under tungsten light, 2848 degrees Kelvin, while 
the samples represented by the curves in Fig. 17C are a color match in 
daylight, about 6700 degrees Kelvin. These curves were supplied by 
F. T. Simon of Sidney Blumenthal and Company. Perhaps the best 
demonstration of a metameric pair that is known to the authors is one 
described by Dexter and Stearns. 48 Dr. Stearns was kind enough 
to provide samples of this pair which were shown under a series of 
five commonly used illuminants when this paper was presented at the 
meeting in Santa Monica. Also shown was the Macbeth Chroma- 
critic which is widely used in the graphic-arts industry for viewing 
colored transparencies. In the Chromacritic the color and intensity 
of the illumination may be changed on a screen to a predetermined 


The authors wish to acknowledge the co-operation of Ralph Far- 
num, Frank Carlson, B. T. Barnes, E. S. Steeb, Jr., of the General 
Electric Company; F. T. Bowditch and W. W. Lozier of the National 
Carbon Company; the British-Thomson-Houston Company; and 
W. D. Buckingham of the Western Union Telegraph Company, who 
supplied information regarding the various lamps described. They 
also wish to thank F. T. Simon of the Sidney Blumenthal Company, 
and E. I. Stearns of Calco Chemical Division, American Cyanamid 
Company, for the metameric samples used at the meeting to demon- 
strate color change of samples with change of light source. 


(1) S. M. Newhall and J. G. Brennan, "The ISCC comparative list of color ; 
terms (1948)," Inter-Society Color Council, 1949. 

(2) I. G. Priest, "A proposed scale for use in specifying the chromaticity of in- 
candescent illuminants and various phases of daylight," J. Opt. Soc. Amer., vol. 23, 
pp. 41-45; February, 1933. 

(3) H. P. Gage and N. Macbeth, "Filters for artificial daylighting, their grad- ; 
ing and use," Trans. Ilium. Eng. Soc., vol. 31, pp. 995-1020; December, 1936 


(4) R. G. Linderman, C. W. Handley, and A. Rodgers, "Illuminants in motion 
picture production," J. Soc. Mot. Pict. Eng., vol. 40, pp. 333-367; June, 1943. 

(5) "SMPE Studio Lighting Committee Report," /. Soc. Mot. Pict. Eng., vol. 
45, pp. 249-260; October, 1945. 

(6) OS A Colorimetry Committee, "Quantitative data and methods for 
colorimetry," /. Opt. Soc. Amer., vol. 34, pp. 633-688; November, 1944. 

(7) E. Q. Adams and W. E. Forsythe, "Radiometric and colorimetric charac- 
teristics of t<he black body between 2800 K and 3800 K," Denison Univ. Bui., J. 
Sci. Labs., vol. 38, pp. 52-68; December, 1943. 

(8) W. E. Forsythe and E. Q. Adams, "Fluorescent and Other Gaseous Dis- 
charge Lamps," Murray Hill Books, New York, N. Y., 1948. 

(9) D. Katz, "The World of Colour," Translated from the German by R. B. 
Macleod and C. W. Fox, Kegan Paul, Trench, Trubner and Company, London, 
England, 1935. 

(10) H. Helson, "Fundamental problems in color vision. 1. The principle 
governing changes in hue, saturation, and lightness of non-selective samples in 
chromatic illumination," J. Exper. PsychoL, vol. 23, pp. 439-476; November, 

(11) H. Helson, "Color tolerances as affected by changes in composition and 
intensity of illumination and reflectance of background," Amer. J. PsychoL, vol. 
52, pp. 406-412; July, 1939. 

(12) H. Helson, "Some factors and implications of color constancy," J. Opt. Soc. 
Amer., vol. 33, pp. 555-567; October, 1943. 

(13) H. Helson and J. Grove, "Changes in hue, lightness, and saturation of 
surface colors in passing from daylight to incandescent-lamp light," /. Opt. Soc. 
Amer., vol. 37, pp. 387-395; May, 1947. 

(14) H. Helson and V. B. Jeffers, "Fundamental problems in color vision. II. 
Hue, lightness, and saturation of selective samples in chromatic illumination," 
J. Exper. PsychoL, vol. 26, pp. 1-27; January, 1940. 

(15) H. Helson and W. C. Michels, "The effect of chromatic adaptation on 
achromaticity," /. Opt. Soc. Amer., vol. 38, pp. 1025-1032; December, 1938. 

(16) D. B. Judd, "Hue, saturation and lightness of surface colors with chromatic 
illumination," /. Opt. Soc. Amer., vol. 30, pp. 2-32; January, 1940. 

(17) R. M. Evans, "Introduction to Color," John Wiley and Sons, New York, 
N. Y., 1948. 

(18) P. J. Bouma and A. A. Kruithof, "Chromatic adaptation of the eye," 
Philips Tech. Rev., vol. 9, no. 9, pp. 257-267; 1947-1948. 

(19) A. A. Kruithof and P. J. Bouma, "Hue estimation of surface colors as in- 
fluenced bv the colours of surroundings," Physica, vol. 9, pp. 957-966; Decem- 
ber, 1942. 

(20) OSA Colorimetry Committee, "Physical concepts: Radiant energy and its 
measurement," /. Opt. Soc. Amer., vol. 34, pp. 183-218; April, 1944. 

(21) C. G., Abbot, F. E. Fowle, and L. B. Aldrich, "The distribution of energy 
in the spectra of the sun and the stars," Smithsonian Misc. Coll., vol. 74, p.* 15; 

(22) P. Moon, "Proposed standard solar-radiation curves for engineering use," 
J. Frank. Inst., vol. 230, pp. 583-617; November, 1940. 


(23) Commission Internationale de L'Eclairage, Proceedings of the Eighth 
Session, Published by University Press, Cambridge, England, 1931, pp. 219-220. 

(24) D. B. Judd, "The 1931 IC1 standard observer and co-ordinate system for 
colorimetry," J. Opt. Soc. Amer., vol. 23, pp. 359-374; October, 1933. 

(25) K. S. Gibson, "Approximate spectral energy distribution of skylight," 
J. Opt. Soc. Amer., vol. 30, p. 88 (A), February, 1940. 

(26) A. H. Taylor and G. P. Kerr, "The distribution of energy in the visible 
spectrum of daylight," J. Opt. Soc. Amer., vol. 31, pp. 3-8; January, 1941. 

(27) N. Macbeth, ST., "Color temperature classification of natural and artificial 
illuminants," Trans. Ilium. Eng. Soc., vol. 23, pp. 302-324; March, 1928. 

(28) W. E. Forsythe and E. Q. Adams, "The tungsten filament incandescent '. 
lamp," Denison Univ. Bui., J. Sci. Labs., vol. 32, pp. 70-131; April, 1937. 

(29) D. B. Judd, "Changes in color temperature of tungsten filament lamps at 
constant voltage," J. Opt. Soc. Amer., vol. 26, pp. 409-420; November, 1936. 

(30) D. Nickerson, "Artificial daylighting for color grading of agricultural 
products," J. Opt. Soc. Amer., vol. 29, pp. 1-9; January, 1939. 

(31) D. Nickerson, "Artificial daylighting studies," Trans. Ilium. Eng. Soc., vol. 
34, pp. 1233-1253; December, 1939. 

(32) D. Nickerson, "The illuminant in color matching and discrimination," 
Ilium. Eng., vol. 36, pp. 373-391; March, 1941. 

(33) D. Nickerson, "Color measurement and its application to the grading of 
agricultural products," U. S. Dept. Agr. Misc. Publ. 580, United States Depart- 
ment of Agriculture, Washington, D. C., 1946. 

(34) D. Nickerson, "The illuminant in textile color matching, An illuminant to I 
satisfy preferred conditions of daylight-match" (a report on Inter-Society Color 
Council Problem No. 13), Ilium. Eng., vol. 43, pp. 416^67; April, 1948; sum- ; 
mary report, /. Opt. Soc. Amer., vol. 38, pp. 458-466; May, 1948. 

(35) F. T. Bowditch, M. R. Null, and R. J. Zavesky, "Carbon arcs for motion 
picture and television studio lighting," J. Soc. Mot. Pict. Eng., vol. 46, pp. 441- 
453; June, 1946. 

(36) F. T. Bowditch and A. C. Downes, "The photographic effectiveness of car- 
bon arc studio light-sources," /. Soc. Mot. Pict. Eng., vol. 25, pp. 375-382; No- 
vember, 1935. 

(37) F. T. Bowditch and A. C. Downes, "Spectral distribution and color-tem- 
peratures of the radiant energy from carbon arcs used in the motion picture indus- 
try," J. Soc. Mot. Pict. Eng., vol. 30, pp. 400-407, April, 1938. 

(38) M. R. Null, W. W. Lozier, and D. B. Joy, "The color of light on the pro- 
jection screen," J. Soc. Mot. Pict. Eng., vol. 38, pp. 219-228; March, 1942. 

(39) R. J. Zavesky, M. R. Null, and W. W. Lozier, "Study of radiant energy at 
motion picture film aperture," /. Soc. Mot. Pict. Eng., vol. 45, pp. 102-108; 
August, 1945. 

(40) G. E. Inman, "Characteristics of fluorescent lamps," Trans. Ilium. Eng. 
Soc., vol. 34, pp. 65-78; January, 1939. 

(41) G. B. Buck, II, and H. C. Froelich, "Color characteristics of human com- j 
plexions," Ilium. Eng., vol. 43, pp. 27-49; January, 1948. 

(42) H. A. Breeding, "Mercury lighting for television studios," Proc. I.R.E., I 
vol. 31, pp. 106-112; March, 1943. 


(43) R. E. Farnham, "An appraisal of illuminants for television studio light- 
ing," /. Soc. Mot. PicL Eng., vol. 46, pp. 431-440; June, 1946. 

(44) W. E. Forsythe, E. Q. Adams, and B. T. Barnes, "Mercury vapor lamps," 
Denison Univ. Bui., J. Scientific Labs., vol. 37, pp. 107-132; August, 1942. 

(45) F. E. Carlson, "New developments in mercury lamps for studio lighting," 
J. Soc. Mot. Pict. Eng., vol. 50, pp. 122-138; February, 1948. 

(46) W. D. Buckingham and C. R. Deibert, "The concentrated-arc lamp," 
/. Opt. Soc. Amer., vol. 36, pp. 245-250; May, 1946. 

(47) W. D. Buckingham and C. R. Deibert, "Characteristics and applications of 
concentrated-arc lamps," /. Soc. Mot. Pict. Eng., vol. 47, pp. 376-399; April, 1947. 

(48) F. C. Dexter and E. I. Stearns, "Example of metamerism," Letter to the 
editor, J. Opt. Soc. Amer., vol. 38, pp. 816-817; September, 1948. 


MR. CARL FREUND: Seeing these different light sources, I think the studio 
make-up department should pay attention to these. I have seen many make-up 
departments where the people are made up under fluorescent light and then set 
under incandescent light. When a Technicolor picture is shown people should 
be made up under proper color lighting. 

MR. LONG: As you increase the pressure on cadmium- mercury lamps, the 
lines consolidate considerably, and in ranges of 20 to 30 atmospheres you have 
almost continuous lines. The cadmium does not come up quite so rapidly, but 
fills in greatly on the lines and appears more like a continuous picture. 

MR. FREUND: I would like to point out one other very important fact in black- 
and-white photography. Usually we make tests of a star before we start to shoot 
a picture usually on a test stage or some place, and measure the foot-candles 
correctly and establish a level for the lighting level. Later in the rushes, and when 
we go into production, usually different lights are used, or newer lamps, and then 
we are supposed to make the close-ups all over again and then everyone says to 
make it just like the test; so we have a chart on what diffusion we use, but we find 
that the skin texture never comes out right. It was different when we shot the 
test than when we actually shot the production there was an older lamp or newer 
lamp in one or the other. But we have a different skin texture, even with the 
proper lamp. 

MR. BELL: Has a photographic test been made and are the effects either more 
pronounced or less than they are here? 

Miss DOROTHY NICKERSON: You get other effects than those just from the il- 
lumination, but I think Mr. Evans can answer it better, because I am sure they 
must have tried it. 

MR. RALPH EVANS: In color photography, we are dealing with the three re- 
ceptive systems somewhat similar to that of the eye, but it isn't expected that a 
pair of colors which match under any illumination should match when photo- 
graphed by any of the known processes. The reason for the match is rather diffi- 
cult to go into briefly, but it is not to be expected that the same colors that match 
under any illumination will match in color photography at the present time. The 
requirement would be that the three receptors have exactly the same properties as 
the television mechanism of the eye, and that at the present time is impossible. 

Color- Order Systems 




Summary The straightforward approach to any color problem requires 
consideration of the properties of illumination, the object, and the receptor. 
Ordinarily the receptor is the eye but in motion picture photography an inter- 
mediate step, the translation of the scene in front of the camera into a scene 
for presentation to the eye, is required. Since a successful final result from a 
color standpoint is a complex combination of many factors, color-reference 
points in the various stages of production are very welcome. 

Many kinds of color-reference material are available and a careful appraisal 
will develop three points of difference. 

One group includes collections of samples which illustrate the color gamut 
of colorants in prescribed mixtures, another group includes samples illus- 
trating systems derived from color-mixture data, and a third group includes 
samples that illustrate systems that deal with various aspects of visual color 

THE FUNDAMENTAL DEFINITION of color can be given as an equa 

Radiant Energy X Visual Process = Color. 

Color is the result of the evaluation of a particular kind of radian 
energy by the visual and related processes of an observer. There ar 
many distinct situations or conditions that give rise to color, bu 
these are all extensions or amplifications of this simple fact. In mos 
cases, the radiant energy has undergone some modification before ar 
observer evaluates it. In fact, many modifications of the radian 
energy from a source or illuminant quite often take place before it L 
finally evaluated by the observer. It is therefore very helpful t 
think of the source of radiant energy as distinct and separate from th 
great variety of modifiers encountered. Thus we may expand om 
equation to the following : 

Source X Modifier X Visual Process = Color. 

In a study of color it may be necessary to use mathematical or othe 
expressions for any part of a single section of this equation, but a 
no time can we ignore the existence and importance of all facton 
included in the equation. 

* Presented May 20, 1948, at the SMPE Convention in Santa Monica. 


It may be in order to repeat these statements in terms of common 
experience. Thus, in a natural outdoor situation radiant energy may 
come from the sun, may or may not be modified by scattering in the 
atmosphere before it is transmitted or reflected by the next modifier, 
which may be any object, and finally reaches the eye of an observer 
who makes an evaluation of color. In an indoor situation the energy 
may come from an electric lamp or other artificial source, may or may 
not be modified by filters or reflectors before it reaches the object for 
further modification, and then through the visual processes of an 
observer is evaluated as color. Ordinarily this is a single sequence of 
events but in photography this sequence occurs twice so that the scene 
hi front of the camera may be translated into a picture for presenta- 
tion to the observer. The picture is a specially prepared modifier 
which allows a representation of the original scene. This sequence 
is shown in the following equations : 

Source X Modifier X Photographic Process = Picture 

Source X Picture X Visual Process = Color. 

The picture resulting from the first equation becomes the modifier in 
the second equation. 

This definition of color includes white, gray, and black as colors, 
and the equation therefore accounts for results of black-and-white 
photography as well as of color photography. Since from a color 
standpoint a successful final result is a complex combination of many 
factors, color-reference points in the various stages of picture pro- 
duction are very welcome. Gray scales and other simple color charts 
are often used in testing the photographic process, and sometimes 
much more complete color charts are needed. The use of such test 
charts is a specialized procedure and so is their production. 

All collections of color chips may be considered as modifiers of the 
radiant energy under which they are viewed or tested. Sometimes a 
single set of reference-color samples may serve a variety of purposes 
while at other times specially prepared material is necessary. 
' Color-reference material is needed by the photographer, designer, 
scene painter, lighting man, costumer, make-up artist and many 
others. This color-reference material usually takes the form of color 
charts in a particular medium such as fabrics, paint, filters, and cos- 
metics. The seeming complexity of the need for so many types of 
reference material has led many people on the search for a complete 
solution to the entire color problem in the form of a single color 

186 Foss February 

There is one fundamental color system (International Commission 
on Illumination) in current use, but it does not provide for a single 
set of reference materials. Color materials may take a wide range 
of forms since factors such as gloss, transparency, texture, and their 
combinations indicate that no one collection of material samples will 
ever satisfy all requirements. 

Because of the special requirements of a particular job at hand, it 
is very helpful to be able to appraise the color collections which are 
available so that the proper or best use of them may be made. 

If, instead of studying available color collections, an attempt were 
made to gather a complete color range by continued assembly of 
existing samples, the result always would be the same, that is, there 
would be too many samples of similar colors, and many regions of 
color space not represented at all. At this point in the assembling 
of such a collection it usually becomes evident that a good series of 
color samples can be better prepared by special production rather 
than by continued assembly. 

The production of a color collection is essentially the manipulation 
of the ingredients in each of the various material forms. The pig- 
ments and dyes in paints, papers, plastics, fabrics, and similar ma- 
terials, are the ingredients whose variation produces the color varia- 
tion and they are known as colorants. There are thousands of 
colorants but many are very similar in color. Color is only one 
property of colorants, and certain properties like stability, com- 
patability, and transparency, often control their applications. In- 
deed, these other properties of the colorants often are the factors 
which must be considered first in producing a collection of color sam- 
ples. Ease of manipulation, and the ability to control ingredients 
in the prepartion of any set of samples will determine whether the 
production of a collection can be completed in a reasonable time. 
This is the principal factor that explains the wide use of paint inj 
making color collections. It is easy to apply paint to a surface, and a 
wide color range is possible with a small number of colorants. It is| 
more difficult to produce specific colors in textiles, plastics, and 
ceramic materials. 

_.It is natural to think of color collections in terms of their material 
ingredients but other important features must not be overlooked. In 
order for a collection of color samples to be called a color system, 
whether the samples constitute the system or merely represent it, it| 
is usually expected that a certain orderliness of interrelation anc 


presentation, as well as wide coverage of color. range, be provided. 
Since the interrelation of the various samples in any single system is 
in part dependent on the geometrical pattern used in its construction, 
the geometry is often the first feature described. It may be thought 
of as the structure which holds the system together. However, the 
geometry of color systems should be kept separate from certain basic 
principles which control the derivation of the actual color samples in 
the collection. These derivation methods are of three general types 
and serve as the basis for a classification of color systems into three 
categories. One group includes collections of samples which illus- 
trate the color gamut of colorants in prescribed mixtures, another 
group includes samples illustrating systems derived by color mixture 
(additive methods), and the third group includes samples that illus- 
trate systems that deal exclusively with visual aspects of color space. 

While at first glance this classification may seem obscure, it may be 
said that until these distinctions are understood, the entire subject of 
color will not be clear. These distinctions deserve emphasis because 
casual inspection fails to discose their importantly different features 

One may, for example, start with color samples representing three 
end points, perhaps black, white, and a strong red. If half-and-half 
mixtures of a particular set of colorants are prepared the result will 
differ greatly from half-and-half mixtures of the same end colors 
mixed additively on disks, and both will differ greatly from color 
samples that are prepared to appear equally spaced between the same 
end points. 

Unfortunately it is not practicable to reproduce in the JOURNAL the 
extensive series of color charts prepared to illustrate the discussion of 
this subject at the Santa Monica meeting. It can be stated, however, 
that all the points discussed are fully capable of illustration. The 
color differences illustrated at the meeting as produced by identical 
proportional treatment applied to the three categories of color systems 
are far too great for them to be considered similar. Wide differences 
exist in the three concepts and they should not be confused or con- 
sidered alike in any way. 

Most color problems call for consideration of these concepts in cer- 
tain combinations and although it is often desirable to describe a 
sample illustrating one concept in terms of another, this does not dis- 
solve the fundamental differences. 

The evaluation of any color system or collection of color chips 
requires recognition of the color concepts it illustrates before attention 

188 Foss February 

is given to the factors of color range, stability, number, size, and form 
of the samples as well as the factors of body and surface properties of 
the samples. 

The geometrical arrangement of the samples and the scales of 
unique variables control the interrelation of the samples in any system 
regardless of the concept it illustrates. 

A knowledge of the three principal derivation methods of color- 
order systems enables one to make the best application of existing sys- 
tems and also indicates how special systems may be constructed to 
satisfy special needs. 

A number of the commercially available collections of color samples 
were described and illustrated at the meeting. In this report we shall 
content ourselves with listing and briefly describing several of the 
better-known systematic collections of color chips, listing each under 
the category into 'which its chief attributes place it. 


Although there are thousands of colorants, various reasons make it 
desirable to use a relatively small number when producing a color 
collection. The material usually thought of as the colorant is seldom 
used alone but almost always in connection with some other material 
whose color is to be altered. The use of a clye on textiles or film pro- 
duces a color range or gamut depending upon the concentration or 
thickness of the colorant on the base whether clear or opaque. Thus 
the color range of a colorant may be thought of as produced by the 
variable ratio of two ingredients. The colorant attends to most of 
the selective modification while the other ingredients attends to the 
nonselective modification. This second ingredient is not ordinarily 
thought of as a colorant but rather as a diluent or extender when 
transparent or as a white when opaque. However, it always con- 
tributes significantly as a modifier. In paint, the colorant (selective 
modifier) is called a toner and is commonly shown in a variable-ratio 
gamut with white (nonselective modifier) to illustrate its principal 
coloration possibilities. 

Series or gamuts of this sort are widely used in all media and the 
general idea is usually the basis for color collections which illustrate 
the color range of the colorants in various combinations. While this 
is not the only useful method it is the one most commonly used. 

Color chips of such systems may be defined in terms of percentages, 
by weight or volume, of the colorants used to make them. 


Comprehensive systems of this sort require representation in three- 
dimensional form. In the construction of colorant-mixture systems 
the choice of component-mixture scales is usually dictated by visual 
considerations. This does not, however, alter the fact that they 
still represent the gamut of the particular colorant combinations. 

Many charts and collections follow the general color gamut method 
and there will be many more. The collections range from extensive 
color coverage made under quantitatively controlled conditions to 
more limited coverage for requirements that are less exacting. 

Each of the systems selected for description is currently available 
and represents a comprehensive collection of color chips based on the 
colOrant-mixture method. Both show the extension gamut of color- 
ants with white and provide formulas for duplicating the color -of the 
samples in the material used in their production. 


In the Plochere color system,* as now available, a small number of 
selected chromatic colorants specially formulated as colors in oil are 
used to produce 26 series of mixed-base paints, each series having six 
steps from a near-neutral to the full chromatic step. Each series of 
six mixed-base paints is formulated to look like a constant-hue series 
and all 26 series of different hues are arranged in the usual sequence 
in radial order around a neutral point. This group of mixed paints, 
26 X 6 = 156, represents the base of a cylindrical color solid which is 
developed by making a white-paint extension series containing eight 
steps from each of the 156 mixed or base paints. The white-paint 
extension scale is a progressive one which was empirically determined 
so that throughout all of the toner gamuts toward white, these series 
give excellent coverage of the color r^nge of the colorants. 

A total of 1248 samples result from this thorough development and 
they are presented as 3- by 5-inch cards in a file box. Each card has 
the formula on the back which shows how each color was made from 
the base paints. The base paints, ten in number, are essential in 
using the formula data given and they are offered for sale to those who 
wish to produce flat wall paint in quantities for average use. 

The Plochere collection was produced to satisfy the demand for a 
relatively inexpensive collection of color cards showing actual formula- 
tions from paints which are currently available. 

* Plochere Color System, G. Plochere, 1820 Hyperion Ave., Los Angeles 27 
Calif., 1948. 

190 Foss February 

The present collection is a simplification and revision of an earlier 
Plochere color guide which used a larger number of colorants. 


The Nu-Hue system of the Martin-Senour Company* is based on a ' 
selection of six chromatic and two achromatic colorants. These basic \ 
paints are specifically designed for interior wall finishes. The system 
consists of one thousand samples presented in the form of a cone, the 
samples arranged in hexagonal closest packing, with planes parallel 
to the base, each plane representing a constant white-paint content. 1 
On the base, the six chromatic colorants are mixed in pairs of neigh- 
boring hues to provide hue coverage, and with black to provide color 
steps toward neutral. The remainder of the base is produced by \ 
combinations of two neighboring hue colorants and black and provides 
271 samples on the base, 54 on the periphery. On this level there are 
nine rings around the black center. 

Each chart above the base has a stated amount of white paint added 
to certain of the base colors, this amount increasing as the levels in- 
crease toward white. Each succeeding higher-level chart has one less 
ring decreasing from nine rings at the bottom to a single white point 
at the peak of the cone. 

In this construction it is possible to show a continuous series from 
any paint on the base to the top, but the samples which occur in each 
series vary widely in number. It is a strict application of the pre- 
scribed mixtures and no deviations are made to include visual consid- 
erations other than variable ratios of the components in the scales. 

The material is available on charts and in a 3- X 5-inch card-index 
file with formulas for each for obtaining a match with the limited 
series of base paints. 

This collection was produced to provide a high-precision formulation 
technique for the production of paint in any quantity to match any 
one of the colors. This is usually done in establishments equipped 
with appropriate mixing devices. 



In" this type of system the color range is determined by the end 
components in additive mixture. Quite often this is expressed in 
terms of the disk-mixture percentages of the components. 

* Nu-Hue Color System, Martin-Senour Co., 2520 Quarry St., Chicago, 111., 1946. 


The general concept provides a structure for the development of a 
wide number of representations of color space. These can follow 
several patterns, as a collection of two-, three-, or four-component 
mixtures, the commonest one being a triangular array using two 
achromatic end points and one chromatic end point. By varying the 
chromatic end point, and repeating the procedure for a considerable 
number of such end points, wide color coverage may be obtained. 
These resulting solids are double conies. 

Several internal geometries may be used, and the two that are 
described below illustrate the internal treatment of the most common 
multiple three-component cases. They have the same external shape 
but different internal co-ordinates. 

The colors within solids determined by strict application of additive 
mixture of the components do not provide the color coverage that 
often is expected. Such deviations have led workers to alter the 
coverage although they still present the samples in a triangular array, 
or some derivation thereof. These altered color ranges can be in- 
cluded in a strictly rigorous additive treatment but would require a 
departure from the commonly used triangular arrangement. 

Although scales used in additive methods are often adjusted to take 
account of visual considerations, this does not alter the fact that they 
still represent color mixture. 


There are 1115 named samples in the Ridgway color charts 1 pub- 
lished in 1912. These charts are still in wide use, particularly in the 
biological and horticultural sciences where the color names used by 
Ridgway have become well known. It is understood that the pub- 
lishers first planned a 5000-copy edition, but that not enough perfect 
color sheets were available to complete this when the first binding was 
made. A supply of enough matching sheets was made later to fill 
in the comparatively few colors (probably not more than 25 or so) 
which were necessary in order to bind more books. While there has 
been a second binding, there has been no really new edition since the 
1912 publication, and the publisher's supply is now exhausted. 

The Ridgway color solid is represented by a double cone with "pure 
colors" at the equator. The upper surface of the solid contains all 
"tints" which are produced by additive mixture with white. The 
lower surface of the solid contains all "shades" which are produced 
by additive mixtures with black. The internal sampling of the solid 

192 Foss February 

follows the same additive mixture paths to white and black from 
desaturated colors produced in turn by additive mixture of the "pure 
color" with a gray midway between the white and black. All of these 
broken colors lie on double-conic solids of successively smaller 

These various series for a given saturation position are shown in 
succeeding sections of the book. The samples of each series are shown 
in vertical order and hue sequence in each section. There is a sys- 
tematic abridgment as the samples become successively more de- 

From inspection there is considerable question as to whether the 
production of the samples followed a strict application of the premises 
of additive color mixture that are so clearly stated by Ridgway in the 
text. There is an indication that these paths tend to follow colorant 
gamuts rather than the disk mixture reported. 

It is understood that botanical workers at the University of 
Toronto* have in process a publication that will provide a conversion 
from Ridgway to Munsell notation. This information will permit the 
careful analysis necessary to establish the actual compliance with color- 
mixture principles. 


The ideal Ostwald system is a good example of a color-mixture 
system. Although it is not possible to obtain the ideal color speci- 
mens by Ostwald for his white and black, or for the semichromes 
specified for his full-color series, his ideal space representation has 
been computed. 

The Ostwald color solid is represented by a double cone with full or 
saturated colors at the equator. The upper surface of the cone con- 
tains what Ostwald calls light clear colors, which are additive mix- 
tures of full colors with white, and the lower surface the dark clear 
which are additive mixtures of full colors with black. The surface, 
except for the selection of the full colors and their angular placement, 
is the same as described for Ridgway. 

Planes of constant Ostwald hue (actually constant dominant wave- 
length) radiate from a central neutral axis, black at the bottom, white 
at the top, each plane having triangular co-ordinates to represent the 
three components, full color, black, and white for any sample on the 

* Professor D. H. Hamly and associates, Botany Department, University of 
Toronto, Toronto, Ontario, Canada. 


plane. Visual considerations dictate the choice of logarithmic scales 
for the mixture ratios and partially control the placement of Ostwald- 
hue planes. 

The hues are numbered beginning with 1 at yellow, either in a 
series of 100, or in the more usual abridgment of 24. The percentages 
of additive white-and-black content are indicated by letters, usually 
a to p, pa being the maximum color in the usual abridgment. The 
notation is written in the order of Ostwald-hue number, white content, 
black content, for example, a saturated red would be written as 7 pa. 

Ostwald himself produced materials in many forms to illustrate his 
system and there are Ostwald color charts included with the author- 
ized English translation of his work. In 1942, however, the Container 
Corporation of America produced in the charts of their Color Har- 
mony Manual 2 what is perhaps the most carefully standardized set of 
Ostwald material that is available. 

These samples were made so that on each Ostwald-hue chart it is 
intended that dominant wavelength be kept constant, and that ICI 
excitation purity be kept constant in the vertical, or "shadow," 
series, with opposite triangles complementary in dominant wave- 
length. The triangle arrangement is maintained, although the color 
range is altered in practice from the additive-color-mixture range to 
conform with the color range of available pigments. 

There are 680 color samples in the Manual. Each is removable 
and is identified by notation. They were prepared by applying a 
pigmented film of appropriate color to a base of clear transparent 
cellulose acetate, thus providing a dull surface on one side, glossy on 
the other. 

ICI specifications, based on spectrophotometric determinations, 
have been reported for this material. 


. Certain methods of producing color collections contain within them- 
selves the possibility of consideration according to different con- 
ceptual points of view. 

A good example of this is the Maerz and Paul collection which is 
made by printing methods and is a combination of color and colorant 
mixture. The half-tone screens over white follow color mixture, but 
the half-tone screens using two colorants may be a combination, since 
in some places they provide an overprinting of two or more inks. As 
a rule, printing processes are combinations of color-mixture and 

194 Foss February 

colorant-mixture principles. It is possible, however, to consider them 
completely as color-mixture systems if the results of colorant mixture 
are taken as components in a color-mixture plan. 

Maerz and Paul 

The Maerz and Paul Dictionary of Color 3 contains several thousand 
colors arranged on pages designed for convenience in printing. Eight 
chromatic inks are used in paired combinations, with screens to pro- 
vide a wide range between the starting points and white. Each is 
printed in a series of 8 charts in which succeeding pages are darkened 
by overprinting with a successively darker transparent gray ink. 
While there are a great many samples on the charts, certain regions 
are represented by very similar samples while larger steps occur 
between successively darker charts. The geometry of the color solid 
which these charts define is awkward in representation and too com- 
plicated to describe here. 

This Dictionary is intended for use as an authority on color names. 
It is based on a wide survey of information regarding names. 


Color collections based exclusively on visual evaluation fall into 
this category. Perhaps the most useful example calls for a uniform 
color space in which each color differs from its neighboring colors by 
some uniform amount. 

There are many ways of sampling this space. One way illustrates 
the three unique attributes in terms of which the system is described. 
Another way could illustrate uniform sampling throughout the space. 
Collections representing these very different methods of sampling 
would have coexistent features. 

The relation of the visual scales depends on the conditions of 
observation. Each collection of samples will be uniformly spaced 
only for the observer, illuminant, and background conditions used in 
its development. 

The first method of sampling is the basis for the only collection of 
samples that has been produced to illustrate exclusively the color- 
appearance concept. No materials are yet available for the second 
type of collection described, although the general principles have been 
reported by the writer, 4 - 5 and certain examples are now in preliminary 
production stages. 



The concept of uniform spacing upon which the Munsell system 6 is 
built is valid for all observers, illuminations, and conditions of ob- 
servation. However, the collection of samples which illustrates this 
system must be differentiated from the concept itself, for it is built 
upon the single set of conditions which describe the average situation 
of normal observer, daylight illumination, and gray background. If 
these conditions are varied significantly, the feature of uniform spac- 
ing may no longer hold for the one set of materials now available. 

The Munsell papers represent a sampling of color space in accord- 
ance with three unique attributes, hue, value, and chroma. Polar 
co-ordinates are used to represent the hue and chroma variables of 
color, and rectangular co-ordinates represent the relations of value 
and chroma. 

The sampling of the solid may be considered as a series of constant- 
hue planes radiating from a central vertical axis, each hue plane 
showing its own range of value and chroma. Another equally useful 
and coexistent plan provides a series of constant-value planes, each 
showing its own range of hue and chroma. Value is uniformly spaced 
in parallel planes for the stated conditions. Hue planes are uniformly 
spaced angularly. Chroma is uniformly spaced cylindrically about 
the neutral axis of the solid. 

The notation for any sample is written in hue, value, chroma 
sequence in terms of the number or letter attached to each scale, for 
example, a medium value, strong chroma red might carry the notation 
oR 5/10. A decimal notation is provided for recording finer dis- 
criminations. About 1000 samples are available that illustrate 
regular positions on the Munsell scales. Many hundred more inter- 
mediate samples are available, usually produced to meet special 
requirements. The colors are available on charts or in disk or sheet 
form cut to various sizes. 

ICI specifications, based on spectrophotometric determinations 
have been reported for most of the Munsell samples. 


In 1947 two color meetings, similar in purpose to the present 
symposium with the Society of Motion Picture Engineers, were 
arranged by co-operation of the Inter-Society Color Council, one for 
the annual meeting of the Technical Association of the Pulp and 
Paper Industry (February, 1947), and the other for the American 

196 Foss 

Ceramic Society, Design Division (April, 1947). Several of the 
papers contained in each series would be of interest in connection 
with the foregoing discussion. Each symposium contains at least 
one paper that discusses directly the subject of color-order systems. 
Bound copies of reprints of each symposium may be obtained by 
SMPE members without cost, as long as the supply lasts, by re- 
quest to the Secretary, Inter-Society Color Council, Box 155, Ben- ] 
jamin Franklin Station, Washington 4, ). C. 

Reference is also made to a discussion of color systems by the 
Colorimetry Committee 7 of the Optical Society of America and to 
three numbers of The Journal of the Optical Society of America, two 
of which 8 - 9 contain a series of papers on the Munsell system, and the 
third, 10 a series of papers on the Ostwald System. 


(1) R. Ridgway, "Color Standards and Nomenclature," A. Hoen and Com- 
pany, Inc., Baltimore, Md., 1912. 

(2) E. Jacobson, "Color Harmony Manual," Container Corporation of 
America, Chicago, 111., 1942. 

(3) A. Maerz and M. R. Paul, "Dictionary of Color," McGraw-Hill Publishing : 
Company, New York, N. Y., 1930. 

(4) Carl E. Foss, "Tetrahedral representation of the color solid," /. Opt. | 
Soc. Amer., vol. 37, p. 529 (A); 1947. 

(5) Carl Foss, "Representations of color space and their applications," Amer. I 
Cer. Soc. Bull, vol. 27, pp. 55-56; 1947. 

(6) "Munsell Book of Color"; standard library and pocket-size editions, 
Munsell Color Company. 10 E. Franklin St., Baltimore 2, Md. 20 hues, 1929; 
40 hues, 1942. 

(7) OSA Colorimetry Committee, "Colorimeters and color standards," /. \ 
Opt. Soc. Amer., vol. 35, pp. 1-25; 1945. 

(8) /. Opt. Soc. Amer., vol. 30, pp. 574-645; 1940. 

(9) /. Opt. Soc. Amer., vol. 33, p. 355-422: 1943. 
(10) /. Opt. Soc. Amer., vol. 34, pp. 353-399; 1944. 

System in Color Preferences* 



Summary Scientific studies of the effects of colors upon people have been 
very limited both in number and in scope. Preliminary investigations offer 
considerable promise that further scientific work along these lines will be 
fruitful. Some examples of results already obtained will be cited to show 
that preferences of people for colors and for color combinations can be pre- 
dicted with satisfactory accuracy. Predictions are possible because there are 
definite relationships between some of the measurable aspects of colored 
objects and the degree of preference for colors as noted for the average 

AN ARCHITECT recently asked, "What do you think is the limit to 
the use of color?" The reply to this was that there seems to be no 
limit. He was referring to the artistic uses of color and my reply was 
in the same vein. From the standpoint of the artistic uses of color, 
alone, recent developments in housing, home furnishing, landscaping, 
motion pictures, and clothing all demonstrate the growing freedom 
and increasing recognition of values in the employment of colors. 
Perhaps this is only one phase of the fact that our culture no longer 
frowns heavily upon sensory enjoyment or of the fact that our civili- 
zation is now past the pioneer stages, and our wealth and leisure foster 
attention to the arts as never before. However this may be, the new 
interest in color has called for the efforts of physicist and engineer, of 
artist and psychologist, and of manufacturer and producer, in at- 
tempts to gain more intimate understanding and improved control of 
colors and also to determine how the consumer is affected by them. 
The role of the psychologist in all this has been to assist in the classi- 
fication of colors according to their appearance and more exclusively 
to learn how people react to color; how well they like single colors and 
color combinations and what kinds of emotional reactions they have 
in the presence of colors. 

It is the subject of likes and dislikes for single colors that will be 
treated Here. Some general conclusions have been drawn from ex- 
periences in the study of this problem. It can be shown that color 
preferences are not a matter of whim, caprice, or of fad; on the 

* Presented May 20, 1948, at the SMPE Convention at Santa Monica. 


198 GUILFORD February 

contrary, they are consistent and orderly. Preferences are related to 
the inherent properties of colors and the preferences of groups of people 
can be predicted from those properties, not perfectly, but with a sur- 
prising degree of accuracy. By "color properties" are meant the i 
three variables that psychologists have generally called "hue" (most 
obviously related to dominant wavelength of the stimulus) ; "chro- 
ma" (color strength, saturation, or richness, most obviously related 
to wave purity) ; and "tint" (the color's equivalence to some point 
or level on the black-white scale). This terminology agrees with that 
of the Munsell system except for the substitution of "tint" for the 
Munsell "value." The reason for this change here is that the author 
wishes to use the term "affective value" to indicate the position of an 
experience on the continuum that runs from pleasantness to un- 
pleasantness. Affective value means degree of preference or degree 
of pleasure. 

The early studies on color preferences suffered seriously from the 
lack of methods for color specification. In spite of this fact, certain 
consistent conclusions were often reached. The chief interest cen- . 
tered in preferences for different hues. The consensus seemed to be 
that the hues blue, green, and red were most preferred, in about that 
order, whereas the hues yellow, orange, and violet were among the 
least preferred. Where discrepancies occurred, there was no way of 
reconciling the results because one investigator's yellow sample, for I 
example, might have differed from that of another investigator, with 
respect to either tint or chroma, or both. 

The relations of preference to tint and to chroma as such received 
almost no attention, although when they did, the consensus seemed to 
be that ligher colors were preferred to darker ones, and the more 
saturated colors to the less saturated. 

My own investigations of these problems began with two hypoth- 
eses: (1) that the affective value of a color is dependent upon all its 
inherent properties, its hue, tint, and its chroma, and (2) that the re- 
lationship of affective value to any one of these properties is a continu- 
ous function, in other words, degree of pleasure increases or decreases 
systematically as hue, tint, or chroma changes in a given direction. 
A serious test of these hypotheses requires intensive study of a liberal 
sampling of colors throughout the color solid. It also requires nu- 
merical specifications of the color sample used and a measurement of 
the degree of liking for each sample. 

To meet the requirement of color specifications, it was decided to 


utilize the Munsell system as the basis. This decision was based on 
the assumption that it is the appearance of the color, not so much its 
stimulus composition, that determines the observer's feelings. This 
assumption may require special investigation. The Munsell system 
seems to offer the evaluations of color most consistent with the 
psychological point of view. To meet the requirement of measure- 
ment of affective value, the judgments of observers were utilized. 
There is insufficient time to go into the details of the technique for this 
here. Suffice it to say that the use of registration of physiological re- 
actions was rejected on the ground that no such indicators of pleasure 
or displeasure have been demonstrated to be either sufficiently con- 
sistent or interrelated to justify their use for this purpose. 

The experimental material consisted of swatches of colored paper 
two inches square.* Surface colors were used because it was felt 
that they represent the most common form of color experience. The 
316 different colors used were sampled systematically from the color 
solid. First, ten alternate hues were selected in the Munsell Book of 
Color. Second, for each hue, samples were sought at alternate levels 
of tint or Munsell value. Third, samples were sought to match each 
chroma level represented. Munsell papers were not available, so it 
was necessary to collect papers from all other known sources. The 
resulting selections only approximated the specifications desired but 
since they could be evaluated in terms of the system, some dis- 
crepancies could be tolerated. Fig. 1 illustrates how the specimens 
were distributed for one of the chosen hues, namely, for red-purple. 

Time does not permit going into details concerning all the experi- 
mental conditions and procedures. 1 Forty individuals of normal 
color vision, twenty men and twenty women, gave their judgments of 
preferences for each color sample on two different occasions. The 
colors were viewed each for a period of five seconds, on a uniform gray 
background rated at N/5 on the Munsell scale, under constant illumi- 
nation. The Munsell specifications had been determined for each 
color under the same illumination. The resulting values for affective 
value of each color were in terms of an eleven-point scale extending 
from to 10, inclusive, with the indifference point (at which experi- 
ences are rated as neither pleasant nor unpleasant) at 5.5. Values of 

* The experimental results referred to in this article were obtained in the Psy- 
chological Laboratory at the University of Nebraska during the years 1935 to 
1939. I am indebted to Ada P. Jorgensen and Patricia C. Smith for material 
assistance in the experiments. 




6 and above therefore represented different degrees of pleasure and 
values of 5 and below represented different degrees of displeasure. 
Having obtained the ratings of affective value for each color, it was 
important to know first whether those values were self-consistent. 
Although there were distinct differences of opinion among individuals 
as to the values of some colors, on the whole agreements far outweighed 
disagreements and the range of values for different colors far ex- 
ceeded the range for a single color. 2 There were some differences of 










o I 




4 6 8 10 


Fig. 1 Diagram of the red-purple plane of 
the Muns"ell color solid, showing by marked 
points the specifications of color samples of 
this hue used in the experiments. 

opinion between the two sexes, on colors of certain hues, particularly, 
so that the data were treated separately for the two sexes throughout. 
Next, let us try to obtain a picture of how the pleasure aroused by 
a surface color varies with each of the properties of that color : first, 
let us isolate each variable hue, tint, and chroma and note its 
unique effects and then study the combined or joint effects of all three 
variables. If we want to discover how preference varies with hue, and 1 
hue alone, we must hold constant the effects of tint and chroma while j 




doing so. Likewise, if we want to know the influence of tint, as such, 
we must hold constant the other variables, hue and chroma. The 
influence of chroma can be ascertained only when hue and tint have 
been held constant. 

Next, let us examine the relation of affective value to tint when 
chroma is zero and there is no hue specification. Fig. 2 illustrates 
the average judgments of both men and women for samples of this 
kind. The curves showing the supposed continuous regression of 
affective value upon tint have been drawn by inspection. Assuming 



234 5 6 7 a 9 


Fig. 2 Points and curves showing the relationship of affec- 
tive value to tint (Munsell value) when chroma is zero. 

that these curves have been properly drawn so as to represent best 
the observed values, which are represented by the points, one can see 
how much those observations differ from the supposed continuous 
trends. Such deviations may be regarded as sampling or experimen- 
tal errors. In spite of their obvious sizes, the hypothesis of continu- 
ity seems fully supported by the results. For achromatic samples, 
under these conditions of illumination, background, and size of sam- 
ple, it appears that, on the average, white samples are neither pleasant 
nor unpleasant, that grays from Munsell value 1 through Munsell 
value 6 are mildly to moderately unpleasant, and the most extreme 




blacks tend to be distinctly pleasant. This kind of relationship, as 
shown in Fig. 2, carries over to many situations in which there is a 
hue and in which chroma is constant at some level other than zero, 
but it becomes modified at higher chroma levels, quite drastically, in 
fact, for colors of certain hues. More will be said on this point later. 
Consider, next, the relation of affective value to hue when tint and 
chroma are constant. Earlier findings, that blue, green, and red are 
preferred to orange, yellow, and violet, are supported, but serious 
qualifications must be made. Not all blues are preferred to all 
yellows, nor all reds preferred to all violets. Fig. 3 shows smoothed 


Fig. 3 Regressions of affective value on hue at chroma level 6 and at various 
constant tint levels when judges were women. 

regressions of affective value upon hue when the chroma is held con- 
stant at 6 and as tint is held constant at various levels. The observed 
points are not shown, in order to avoid congestion in the illustration. 
Their discrepancies from the curves shown are comparable with those 
exhibited in Fig. 2. Smoothing had been done not only in the plane 
of this particular figure but also in planes of constant hue and con- 
stant tint, until we arrved at consistent results in all planes. Inter- 
polated values thus became possible and a more complete picture can 
be given. In Fig. 3, not all the curves are complete, either because 
colors do not exist for some of the combinations of hue, tint, and 
chroma implied by the figure or because specimens and observations 




fell short of complete sampling of those that do exist. Similar figures 
have been drawn representing other chroma levels. Fig. 4 shows 
what happens at the chroma level 2. There is consistent evidence 
that the judgments of women were more sensitive to variations in 
hue than were those of men. Women will not be surprised to hear of 
this result. 

Figs. 3 and 4 illustrate several principles. Continuity is again 
obvious. The points of maximal preference for each combination of 
constant tints and chromas come in about the same positions on the 
hue scale near red, green, and blue. The minimal preferences come 

Fig. 4 Same as Fig. 3 except that chroma was constant at level 2. 

near yellow, violet, and, in some instances, between green and blue. 
The general level of any particular curve reflects clearly the influence 
of tint and the range of affective values for any curve reflects rela- 
tively more the influence of chroma. The influences of these variables, 
as was said before, are best noted when hue is constant, however, and 
we return to these variables later. Here we are primarily interested 
in the effects of hue. In another publication, 3 the author suggested 
that curves such as those in Figs. 3 and 4 can be regarded mathemati- 
cally as periodic functions for which empirical equations could be 
written. A Fourier analysis of some data of this kind tended to show 
that the chief components of one of these curves, when chroma is 




approximately 8 and tint is appromimately 6-, are the first and third 
harmonics. There are interesting implications of this finding, but 
the use of such involved mathematical equations for predicting color 
preferences is not recommended ; a much more practical method will 
be demonstrated. 

Fig. 5 shows the regressions of affective value upon chroma when 
the hue is yellow and when tint is held constant at different levels. 




Z 4 6 6 10 12 14 


Fig. 5 Regressions of affective value on chroma 
for colors of yellow hue at different fixed tint levels 
when judges were men. 

These curves are rather characteristic of those obtained for most hues. 
Earlier studies had brought out the general conclusion that affective 
value increases as chroma increases. We see in Fig. 5 that this is not 
true throughout the range of chroma. Near zero chroma, affective 
value usually decreases as chroma increases, reaching a minimum 
level when chroma is in the region of 2 to 4. It has been suggested 
that this may be caused by the fact that at such low chromas the 
nature of the hue is difficult for most observers to identify. This 
arouses annoyance and hence displeasure. However, this may be, for 




most hues at most tint levels the minimum preference is not at zero 
chroma but at some degree of saturation greater than zero. There 
was one noteworthy instance in which preferences decreased again 
for high chroma levels. This was for the hue purple-red, in judg- 
ments of women only. There was decreasing pleasure for colors 
above chroma 10 and above Munsell value 4. Whether this reflected 
a temporary fad or is a more fixed disposition of women is unknown. 
It should be said that there may be certain individuals who have a 




Fig. 6 Regressions of affective value on tint for a typi- 
cal "warm" color at different fixed chroma levels when 
judges were men. 

general preference for unsaturated colors and a dislike for "loud" or 
saturated colors. Some psychologists have maintained that there 
are two types of individuals with respect to reactions to saturated 
colors; one type likes them, the other does not. The results re- 
ported to you pertain to small colored surfaces and represent the 
average judgments of a number of supposedly normal observers. 
One should not generalize too far from these results. There is an- 
other finding in the psychological literature to the effect that unsatu- 
rated colors gain relatively more by increased size of stimulus whereas 




saturated colors gain less, and in fact may lose in affective value. 
This problem is very important and must be solved systematically 
and in thorough fashion before laboratory findings such as those re- 
ported here can be carried over to practical application. 

The regression of affective value upon tint when hue is constant, and 
when chroma is constant and not zero, is illustrated by Figs. 6 and 7. 
It is necessary to show two illustrations for this because two groups of 




5 6 


Fig. 7 Similar to Fig. 6 except that the results are for a 
"cool" color as judged by women. 

colors yield different effects in this respect. The two groups may be 
roughly called the "warm" and the "cold" colors. The former tend 
to yield a regression with the lower preferences at moderate tint 
levels. This tendency is consistent with the relation of affective 
value to tint when chroma is zero (mentioned earlier), and holds 
generally for colors whose hues are yellow, red, or purple. (See Fig. 
6.)- For the "cool" colors, on the other hand, the higher preferences 
may appear at moderate tint levels. At least, in most "cool" colors 
the samples of moderate tint are relatively more agreeable. This is 
true for the greens and blues, including violet-blue. A general princi- 
ple which seems to embrace both types of regressions is that colors tend 




to be most preferred at tint levels at which they can be most saturated. 
The yellow-reds and yellows can be most saturated at the higher tint 
levels and at those levels one finds the most agreeable colors at each 
degree of chroma. The greens and blues can be most saturated at 
some moderate or low tint level, and, for any fixed degree of chroma, 
that is where the maximal preferences occur. An outstanding ex- 
ception to this principle comes in the region of red and purple which 
can be most saturated, also at moderate or low tint levels and yet they 
yield results similar to those in the region of yellow. 





Fig. 8 An isohedon chart showing on the Munsell 
, plane for yellow the lines of equal preference as 
judged by the average man. 

From what has been briefly reported thus far, it can be seen that 
the relationships of preferences to hue, tint, and chroma are so con- 
tinuous and systematic that, knowing the specifications of a color 
sample, we should be able to make a fairly accurate prediction of its 
affective value. With series of charts such as those already shown, 
one could start with the three Munsell designations for a color and 
look up its probable average affective value. One might even set up 
empirical equations in which the three values could be substituted 




and the equations solved to predict the expected affective value. 
This approach, as well as that using the charts of the types illustrated, 
however, would be too unwieldy for practical purposes. As a much 
more practical substitute, charts like those illustrated in Figs. 8 and 9 
have been prepared; one for each of the ten alternate hues in the 
Munsell system. Twenty such charts could be prepared, but pref- 
erences for the hues not represented in the ten could be found, by 









Fig. 9 Same as Fig. 8 except the plane is for Munsell 
blue and the judgments are for women. 

Each chart, representing a constant hue, shows the lines of equal 
affective value in steps of one-half unit. By analogy to weather maps, 
the lines have been called "isohedons" (lines of equal pleasure) and 
the diagrams have been called "isohedon charts." Separate charts 
have been drawn for men and women, for, although similarities be- 
tween reactions of the two sexes to colors generally outweigh differ- 
ences, in the charts, for some hues in particular, the differences are so 
large that much accuracy in prediction is to be gained by recognizing 
those differences. The greatest sex differences occur in reactions to 


the hues of red-purple, purple, and yellow-red. Greatest similarity 
between the sexes occurs in the region from yellow to blue. For 
either sex taken alone, the predictions are most accurate when hues 
are red, yellow, green, or blue, and poorest when hues are purple, 
yellow-red, or green-yellow. The last statement suggests that 
familiarity with primaries may be a factor favoring stability of judg- 
ments. The over-all accuracy of prediction can be expressed in 
several ways. From one point of view, we can say that the predictions 
are nearly 80 per cent accurate for the average judgments of the 
women, and at least 85 per cent accurate for judgments of the men. 4 
The men who read this will not be surprised at the somewhat lower 
predictablity of the reactions of women. 

Having noted statements regarding the accuracy of predictions of 
average affective values from the known specifications of color 
samples, let us- be ready to recognize their limitations and to qualify 
the conclusions. First, it must be emphasized that we are talking 
about averages and not about single reactions in single individuals at 
particular moments of time. Fortunately we are usually called 
upon to make predictions of reactions of masses of individuals over 
periods of time, or of average people, and not of isolated reactions. 
The predictions hold under a set of conditions such as prevailed when 
these experimental results were being obtained. They are based upon 
the assumption that all other determiners of color preferences are held 
constant. They should hold for preferences of colored surfaces, of a 
given size, on a given background, under a given illumination, and 
for a population of individuals similar to those who rendered their 
judgments in the experiments. Before we are justified in predicting 
the preferences for groups of other kinds, with colors of different 
sizes, on different backgrounds, attached to different objects cloth- 
ing, houses, furnishings, and cars and under different illuminations, 
much further experimental study will be needed in order to deter- 
mine how these other factors influence preferences. It is quite possi- 
ble that these other factors do not entirely overcome or even over- 
shadow the effects of the color properties themselves as determiners 
of pleasure derived from colors, but this needs to be demonstrated 

Problems of the effects of color use and o*f color combinations are 
of even greater practical importance than the problems of preferences 
for single colors. These more practical problems cannot be fully 
solved, however, until those pertaining to single colors are also 


solved, and the most economical approach to these problems in 
the long run will be through the study of effects of single colors. It 
has already been shown experimentally that there is a strong relation- 
ship between the pleasure aroused by a color combination and by its 
components taken alone. Principles which hold for the preferences 
for single colors should also have an important bearing upon prefer- 
ences for combinations of those colors. Although uses of color are 
very strongly determined by cultural and conventional forces, the 
range of these effects is undoubtedly limited by the influence of the 
properties of the colors themselves. 

In conclusion, it may be said that a systematic study of preferences 
for colors, when the entire color solid is thoroughly sampled, shows that 
there are definite relationships between the degree of pleasure that a 
color arouses and the intrinsic properties of the color itself. As hue, 
tint (Munsell value), and chroma change continuously in a fixed direc- 
tion, judgments of pleasure also change accordingly. The relation- 
ships are continuous but not simple. The relation of affective value 
(preference or pleasure) to any one property of color is modified by 
changes in any other property, but in a systematic manner. When 
other factors are held constant, including size, illumination, back- 
ground, and type of observer, a prediction of average affective value 
can be quite accurately made from the knowledge of Munsell specifi- 
cations of the color sample. By conversion from Munsell specifica- 
tions to those of other systems, presumably just as accurate predictions 
of the same kind could be made. For fully useful predictions in 
practice, other determiners will need to be taken into account, de- 
terminers that were held constant in obtaining the results reported. 
The results already obtained suggest considerable promise for the 
fruitfulness of further systematic research on other determiners of ; 
liking for colors and also for their different uses and combinations. ' 
Such research should be amply rewarding for those who desire the 
solution of color problems as they affect the human observer. 


(1) More of the experimental details are described in an article, J. P. Guilford, 
"A study in psychodynamics," Psychomelrika, vol. 4, pp. 1-21; 1939. 

(2) Information on the reliability of the data is presented in an article, J. P. > 
Guilford, "There is system in. color preferences," J. Opt. Soc. Amer., vol. 30, pp. \ 
355-359; 1940. 

(3) J. P. Guilford, "The affective value of color as a function of hue, tint, and [ 
chroma," J. Exper. Psychol, vol. 17, pp. 342-370; 1934. 

(4) For additional information on this point, see reference 2. 

16-Mm Release Printing Using 
35- and 32-Mm Film* 


Summary This paper describes the method now used by Paramount in 
making 16-mm release prints from 35-mm original studio productions. The 
purpose of this method primarily is to produce 16-mm release prints com- 
parable to the 35-mm sound and picture print quality and standards. A con- 
siderable advantage is gained by utilizing standard 35-mm facilities such as 
developing machines, rewinds, take-ups, and splicing equipment for most of 
the operations. Special 35-mm width films with 32-mm symmetrical 16-mm- 
type perforations are used for the sound and picture release negatives, with 
two tracks on each film so that two reels of 16-mm release are obtained with 
each developing and printing operation, thereby saving valuable time and 

A specially designed optical-reduction printer is used to make the double- 
track picture release negative. A specially designed sound recorder is used 
to produce the highest possible quality of re-recorded sound negative. A 
specially designed contact release printer is used to print the 35-mm width, 
32-mm perforated sound and picture double-track negatives to the 32-mm 
width fine-grain release positive stock. Twelve-hundred-foot rolls of the 
32-mm print stock are used, corresponding to 3000 feet of the original 35- 
mm production. The 32-mm film is developed in a developing machine 
modified for this width and the finished print is then slit to make the two 
16-mm release prints. 


IN CONSIDERING the problem of 16-mm release from 35-mm feature 
productions, it was concluded that the most economical and high- 
est quality operation should include: 

(a) Electrical re-recording of the 35-mm release sound track to 
permit making the necessary changes in frequency response and com- 

(b) The use of 35-mm width film with double 16-mm-type .per- 
forations for both the release sound and release picture negatives in 
order to utilize existing 35-mm film-developing and -handling facilities ; 

* Presented April 22, 1947, at SMPE Convention in Chicago. 211 





(c) The re-recording to two sound tracks on the release sound 
negative and the optical printing of two picture tracks or images on 
the release picture negative to save handling time in development. 
The sound and picture tracks run in opposite directions on the two 
sides of the film; 


1 1 35 mm Width Film 
| | 35mm Sprockets 


|L____jj 32mm Sprockets 

Std 35mm PRINTS 

{n $ 32mm Width Film 
_Jij 32mm Sprockets 

C. P. Contact Print 

Operation. ST| 





CP / 


CP 5 




ND Optical Reduction: 



r- ' ^^Tl 


j r 35-32 mm DUPE 1 j 
' ND , | DOUBLE 







^ ~~H 


Operation | [^."ji'm^, DOUBLE" 

L __1 1J 1 

Split Down 

\ X' 

J r 32mm DOUBI 

,E 1 1 
ACKl 1 

Fig. 1 New 16-mm film processing sequence (dashed 
lines), and conventional 35-mm regular and dupe process- 
ing (solid lines). 

(d) The use of 32-mm release print stock to accommodate two ] 
complete 16-mm movietone prints. This film is split after develop- 
ment for assembly on 16-mm reels. 



Fig. 1 shows a block diagram of the various film steps involved in j 
making (a) normal 35-mm release, (b) duped 35-mm release, and (c) i 
the new operation for 16-mm release. Solid blocks show existing 1 


35- AND 32-MM FILM 


35-mm operations while the three blocks with dashed outlines indi- 
cate the added operations required for 16-mm. The following special 
films are used : 










35-Mm Negative 
A 1.378 + 0.0000, -0.0040 
B 0.131 0.0020 
C 0.056 max 
D 0.028 max 
E 0.060 0.0010 
F 0.300 0.0005 
G 0.004 max 
H 0.402 0.0020 
I 0.050 0.0004 
J 0.072 =*= 0.0004 

R 0.010 

32-Mm Positive 
1.26 + 0.0000, -0.0040 
0.072 0.0020 
. 056 max 
0.028 max 
0.060 0.0010 
0.300 0.0005 
0.004 max 
0.402 0.0020 
0.050 =*= 0.0004 
0.072 =*= 0.0004 

The dimensions C, D, and .# are those used by Para- 
mount in recording and printing for 16-mm release: how- 
ever, the track may be as wide as 0.080 (upper section of 
drawing), in which case C becomes 0.036 and D becomes 
0.018. All dimensions are in inches. 
Fig. 2 

The "35- to 32-Mm Double-Sound-Track Release Negative" is 
made by electrical re-recording from the sound track of a regular 
35-mm composite release print. 

The "35- to 32-Mm Dupe Double-Picture-Track Release Negative" 
is made by optical reduction from the picture on the already existing 
35-mm master fine-grain positive picture. 

The "32-Mm Double-Composite-Track Release Print" is made by 
contact printing, first from the 33- to 32-mm sound negative and then 




from the 35- to 32-mm picture negative. This is practically the same 
method, practice, and operation of making all present 35-mm stand- 
ard release prints. 


High-quality, 16-mm release recording requires different frequency 

response, level, and compression characteristics than 35-mm release. 

Therefore the use of optical reduction of the sound track was ruled 

out. Since electrical re-recording was indicated, a decision had to be 





Film dummy 12 

Dummy amplifier 13 

Film equalizer 14 

Mixer pot 15 

Line amplifier 16 

Standard release recorder 17 

Bridging coil 18 

Band-pass filter 19 

Attenuators 20 
Compressor amplifier 

High-frequency equalizer 
Recording amplifier 

Phototube amplifier 
Direct-monitor bridging coil 
Monitor relay 
Monitor equalizer 
Monitor amplifier 
Monitor horn 

Fig. 3 16-mm sound-recording channel. 

made between the use of 16-, 32-, or 35-mm width sound re-recording 
negative. As it was desirable to use standard 35-mm negative de- 
veloping machines and accessory-handling equipment, 35-mm width 
film was selected as shown in Fig. 2, with 32-mm symmetrical perfora- 
tions to tie in with the printing of the final release track. The extra 
1.5 mm of film width G is located symmetrically on the outside edges 
of the film. This extra film is never removed, nor is this film ever 
split, which permits the use of 35-mm width nitrate films for the 35- 
to 32-mm release sound and picture negative tracks. 

Re-Recording Channel 

The sound pickup for 16-mm re-recording is made from a regu- 
lar 35-mm release print, using a standard re-recording dummy, mixing 
portion, and monitor horns of the normal 35-mm re-recording channel. 

Fig. 3 shows the block schematic of the re-recording system used. 
The over-all frequency characteristic of the entire 35-mm portion of 


35- AND 32-MM FILM 


this channel, including a reproducer (1), reproducer amplifier (2), 
film equalizer (3), mixer (4), amplifier (5), the 35-mm film recorder 
(6), and the processed film is nearly flat as shown by curve A of Fig. 
4. When re-recording for 16-mm, the film equalizer (3), in Fig. 3, is 
reduced in equalization to produce an over-all 35-mm channel charac- 
teristic shown by curve B in Fig. 4. The 16-mm portion of the re- 
recording channel is connected to the bridging bus of the 35-mm chan- 
nel by the bridging coil (7) (Fig. 3). The regular monitor amplifier 
(19) and horn (20) of the 35-mm channel are connected to the direct 
and phototube-monitor outputs of the 16-mm channel as shown. 



Fig. 4 Characteristics of 35-mm and 16-mm re- 
recording channel. (A) 35-mm over-all, including film, 
(B) 35-mm over-all, with reduced film equalization, (C) 
16-mm high- and low-pass filters, (Z>) 16-mm film equalizer, 
(E) 16-mm monitor equalizer. 

In accordance with the general suggestions of the Society of Motion 
Picture Engineers and the Academy of Motion Picture Arts and 
Sciences, the frequency characteristic of the 16-mm channel is re- 
stricted by high- and low-pass filters as shown by curve C in Fig. 4. 

An RCA Type MM0206C compressor-amplifier (10) (Fig. 3) pro- 
vides the necessary compression and channel gain. This amplifier is 
equipped with a de-esser equalizer in the rectifier network to increase 
the compressor action for high-frequency speech components. The 
limiting characteristics of this amplifier at 400 and 4000 cycles are 
shown in Fig. 5. Since the film equalizer of the 35-mm channel is 
operating at reduced equalization to minimize high-frequency opera- 
tion of the compressor, it is necessary to follow the compressor- 
amplifier by a 16-mm-type high-frequency equalizer (12) (Fig. 3) to 
obtain the desired over-all characteristic. This added equalizer 
characteristic is shown by curve D in Fig. 4. 




The over-all characteristic from the input of the 35-mm film equal- 
izer to the input of the 16-mm modulator is represented by Fig. 6. 
It should be noted, however, that because of the variable-level and 
frequency-compression characteristics of the compressor-amplifier 
both ends of this characteristic are subject to change since the curve 
shown was obtained for a single level input, without compression, i.e., 
the summation of curves B, C, and D in Fig. 4. 



Fig. 5 Compressor-amplifier characteristics at 400 and 4000 cycles. 

The 16-mm direct monitor circuit is obtained through the bridging 
coil (16) in Fig. 3. Since there is considerable high-frequency droop 
in 16-mm reproducers, the monitor characteristic of this channel is 
also reduced by the monitor equalizer (18) which has the characteris- 
tic shown by curve E of Fig. 4. 

Recording Machine and Modulator 

The chassis of a 35-mm Western Electric D-86715 recorder was 
used as the basis of the special recording machine. The sprocket 
diameters were reduced in size in order to reduce the film speed from 
90 feet per minute to 36 feet per minute. The filtering was read- 
justed for excellent flutter characteristics at this speed. Standard 
35-mm width rollers were used but special sprockets had to be in- 
stalled for pulling this special film which has 16-mm-type sprocket 
holes as shown in Fig. 2. The film magazine was specially designed 
so that it could be reversed on the machine, permitting the film to be 
run through the recorder first in one direction and then the other 


35- AND 32-MM FILM 


without rewinding. The magazines were built to handle 2000 feet 
of film anticipating the future use of 1200- to 1600-foot rolls of 
negative stock. 

The modulator is a modified Western Electric Type RA-1152 which 
produces variable-density intensity-type modulation. An equalizer 
used with this modulator gives the valve a flat amplitude-frequency- 
response characteristic over the desired range - A standard RA-1124 
noise-reduction amplifier is used in the conventional manner, with 10 
decibels noise reduction and 6 decibels margin. No reverse bias is 

-555 icsr 


Fig. 6 16-mm-channel characteristic. 

A modification of this modulator permits the recording of a sound 
track on either side of the film centerline without readjustment of the 
lamp, modulator, or film drive. In order to do this, the long-filament 
lamp normally employed with 200-mil push-pull recorders was used, 
together with one side only of an RA-1061, four-ribbon, push-pull- 
type valve, the other side being permanently masked off. Since the 
centerlines of the lamp, optical system, valve septum, and film coincide, 
it is only necessary to turn the valve end for end on the magnet to re- 
verse the position of the sound track on the film. For the black-and- 
white film operation described in this paper, only one position of the 
valve is required. The ability to reverse the valve, however, may be 
needed in the future in connection with reversal or color-film operations. 

The sound tracks are recorded near the center of the film as shown 
in Fig. 2, which permits the cylindrical lens of the modulator to be 
installed close to the, film, between the recording sprocket teeth. The 
60-mil width sound track is used for release. Masking exists in the 
light valve so no further masking is required in the printer. 

The RA-1152 modulator is equipped with a high-quality deflector- 


type phototube monitor; therefore excellent monitor quality is ob- 
tained. This is an asset in this type of re-recording work. The 
excess illumination resulting from the 40 per cent reduction in film 
speed was reduced by using partially silvered deflector glass which 
attenuated the light to the film and at the same time the reflected 
light to the phototube was increased. 

Re-Recording Synchronizing Procedure 

The present practice is to re-record each reel twice on the same piece 
of negative film, once in one direction on the negative and the second 



Fig. 7 32-mm release print. 

time in the reverse direction. The start synchronizing mark on the 
35-mm master film, from which the re-recording is to be made, must 
be at the standard Academy distance from the first splice and an end 
synchronizing mark is placed at an equal distance from the end splice 
of each reel. The unexposed negative must carry both a start and 
end synchronizing mark corresponding to the master track. In re- 
recording the start mark is placed on the unexposed negative, the 
film is run through the recorder in one direction, an end synchroniz- 
ing mark is placed on the film, the magazine is reversed without re- 
winding, the end mark is then used as the start synchronizing mark 
for the second re-recording as the same reel is made to run through 
the recorder in the reverse direction. (See Fig. 7.) 

This procedure of double re-recording of the same reel is used to 
facilitate the release printing and to avoid film losses which other- 
wise would exist because of differences in reel length. The same emul- 
sion-type film is used for the re-recording negative as is used for regular 
35-mm release negative, the only difference being hi the type of per- 
foration employed. Therefore, the same release sound negative de- 
veloping machines and solutions can be used for both 16^nm and S5-mm 


In order to reduce the picture optically from the 35-mm fine-grain 


35- AND 32-MM FILM 


master positive to the special 35- to 32-mm dupe picture negative, a 
specially designed daylight optical-reduction printer was manufac- 
tured by the Bell and Howell Company (Fig. 8). The components 
are 1, the main machine pedestal; 2, an accurately machined base 
plate; 3, complete lamphouse assembly with small-size exhaust fan; 
4, a light-shielding tube; 5, the 35-mm intermittent mechanism to ac- 
commodate the master positive; 5A, the 35-mm master positive 

Fig. 8 Step optical reduction picture printer. 

take-up; 6, critical objective lens adjusting device; and 7, a standard 
35-mm camera head modified to accommodate 35-mm width negative 
with double 16-mm perforation pitch. 

In two operations, the two picture tracks of reel 1 A, for example, are 
laid down on this film in the positions shown by the dotted lines in 
Fig. 2, running in opposite directions like the sound track. The off- 
sets of the picture tracks longitudinally on the film are shown by Fig. 




7, which illustrates one reel of release print before splitting. While 
the sound tracks are exactly opposite each other, with the "start" 
synchronizing mark of one recording corresponding in position with 
the "end" synchronizing mark of the other recording, the picture 
tracks are offset 26 frames on each side so that the film will play in syn- 
chronization on a 16-mm reproducer. 

Fig. 9 Modified printer head for printing double 32-mm 
picture or sound. 

The 35- to 32-mm dupe picture negative is the same type of emul- 
sion as used for the regular 35-mm dupes but has of course the special 
32-mm perforations. The regular 35-mm dupe-negative developing 
machine and developer are used which represent an additional econ- 
omy in operation since no new equipment of this type need be installed. 

Reels 1A, IB, and 2A of both picture and sound negatives are as- 
sembled eliminating identifications between reels 1A and IB and be- 
tween reels IB and 2A thereby permitting the printing of three regular 


35- AND 32-MM FILM 


release reels in proper continuity for mounting on standard 1200-foot 
16-mm reels after slitting. 


The sound and picture 35- to 32-mm negatives are contact-printed 
to the 32-mm width standard fine-grain release positive using a spe- 
cially designed Bell and Howell continuous printer (Fig. 9) with a modi- 

Fig. 10 

fied Model D printer head for printing picture or sound from a 35-mm 
width double-16 perforated negative to a double-16 perforated 32-mm 
width standard fine-grain release positive, which is slit after develop- 
ment. The sprocket holes are of the same type as on the 35- to 32-mm 
negatives but this print stock is only 32 mm wide. Therefore, this 
printer uses guide rollers for film positioning which keep the center- 
lines of the two films together within 1 mil which is sufficiently ac- 
curate for release purposes. Both sound tracks on the sound nega- 
tive are printed at the same t^rne as are both picture tracks on the 
picture negative when it is run through. 



This double-track composite print is developed in standard release 
positive developer in a 35-mm-type developing machine which has 
been slightly modified to handle the narrower 32-mm width film. 
This is the only modified developing machine used in the entire opera- 
tion. Please note, however, that the modifications are such as to 
permit 32- or 35-mm width film development without interruption of 
the continuous operation. 


After development, the double-track composite print is run through 
a specially designed slitter which splits the film down the center as 
shown in Fig. 10. The weave of this slitter is held within 1 mil. The 
film is then ready for final assembly. Since 1200-foot rolls of release 
positive are used, they represent three 1000-foot rolls of the original 
35-mm production. Synchronizing marks are removed between rolls 
and the film spliced together in slightly less than 1200-foot shipping 


The complete planning and method of handling the 16-mm negative 
and release project was suggested and worked out by Frank La Grande; 
and from his specifications, the optical-reduction printer, the release 
printer, and slitter were engineered and manufactured by the Bell and 
Howell Company. 

The special recording machine and facilities for re-recording and ob- 
taining the sound negative were worked out by Bruce Denney and C. 
R. Daily under the supervision of Loren L. Ryder, all of the Holly- 
wood Paramount Studio Sound Department. 


The procedure described in this paper is now being used by Para- 
mount in the United States and is also being installed in the new 
Paramount Laboratory in London, England. Mr. La Grande has 
submitted this procedure to the British Kinematograph Society for 
general acceptance. 

With improved techniques of optical reduction it is also possible 
to obtain the sound negative by the optical-reduction method in 
place of sound rerecording. This presupposes that the frequency 
characteristic and volume range on the 35-mm sound record are cor- 
rect for 16-mm reproduction. If the original recording is a density 
recording, printer-light changes can be made as a part of the duping 
process in which case some reduction in volume range may be 

Three Proposed American Standards 

PROPOSED American Standards for cutting and perforating 
J_ 32-mm film appear on the following pages. They have been de- 
veloped by a subcommittee on film dimensions, of the SMPE 
Standards Committee. 

Film of this type has been used since 1934 although there never has 
been a formal standard. During the intervening years a number of 
changes have been made in the dimensions. Debrie, who was the 
originator of this film, was aware that the slitting of the 32-mm film 
into two 16-mm widths might be inaccurate. This inaccuracy would 
make one half wider than the other half, and might cause trouble 
since the wide half might stick in the projector gate. Therefore, he 
made the original French film narrower than twice the width of 16- 
mm film. The first French film was about 1.252 inches in width. 
Manufacturers in this country made film of this width for some time 
but later widened it to 1.257 inches, an increase of 0.005 inch. 

It appears that there have been four or five slightly different styles 
of perforating in use at various times. The values currently adopted 
for the width of the film and for the transverse pitch of the perfora- 
tions are believed to be acceptable to all manufacturers. The differ- 
ences between the present standards and the earlier dimensions are so 
slight, it is doubtful that the users can perceive them. The dimen- 
sions of the perforation, the longitudinal pitch, and the like, are the 
same as that of the current 16-mm film and the dimensioning of the 
drawing is in keeping with those standards (Z22.5-1947 and Z22.12- 

It will be noticed that the new standards include one for 35-mm 
film with 32-mm perforations . The reason for the existence of this film 
is that' it can be processed on 35-mm sprocketless developing machines 
with consequent saving in equipment. This film is commonly used 
for sound recording and reduction negatives. The negative thus 
made is printed in the usual fashion. In general, this 32- on 35-mm 
film is not used for release purposes. However, the fact that people 
other than manufacturers can perforate 35-mm film in this way has 
led to some concern. If 35-mm nitrate film were to be perforated 
with 32-mm perforations, it might later be slit to 16-mm size and be 
used in projection equipment. Therefore, the standard includes a 
proviso, "This film should not be made on nitrate base because if this 


material were slit to 16-mm it might be used on a projector with conse- 
quent danger of fire." 

No proviso of this sort has ben indicated in other standards for the 
reason that it is an unwritten law in film-manufacturing companies 
that no nitrate base should ever be slit to 8-, 16-, or 32-mm widths. 
The manufacturers do, however, slit both nitrate and acetate film 
to 35-mm dimensions. Other film users sometimes buy unperf orated 
film and perforate it as they see fit. It was thought, therefore, that 
special attention should be called to the danger that might result if 
nitrate film were perforated to any dimensions that might make it 
usable on 16-mm projectors. 

These proposed standards are being published for a ninety-day 
period for your comment and criticism. If no adverse comment is 
received before the end of this period, these proposals will be sub- 
mitted to the Standards Committee for final approval. 

Letter Symbols for Physics 

Another new standard that will be of interest to many motion pic- 
ture engineers is ASA Z10.6-1948 "Letter Symbols for Physics" re- 
cently announced by the American Standards Association. 

The standard suggests that authors who are preparing manuscripts 
give careful attention to the use of symbols which should always be 
clearly defined to avoid errors in interpretation. "Letter symbols are 
to be distinguished from abbreviations, mathematical signs and opera- 
tors, graphical symbols, and chemical symbols : 

(a) Abbreviations are shortened forms of names and expressions employed in 
texts and tabulations and should not be used as symbols in equations. 

(b) Mathematical Signs and Operators are characters used with letter symbols 
to denote mathematical operations and relations. 

(c) Graphical Symbols are conventionalized diagrams and letters used on plans 
, and drawings. 

(d) Chemical Symbols are letters and other characters designating chemical 
elements and groups." 

Copies are now available from the American Standards Association, 
70 East 45 Street, New York 17, N. Y., at the price of $1.00. 




Proposed American Standard 
Cutting and Perforating Dimensions for 

32-mm Sound Motion Picture 

Negative and Positive Raw Stock 

December 1 948 

Page 1 of 2 Pages 










1.257 0.0.01 

31.93 0.025 


0.300 0.0005 

7.620 0.013 


0.0720 0.0004 

1.83 0.01 


0.0500 0.0004 

1.27 0.01 


0.036 0.002 

0.91 0,05 


Not > 0.001 

Not > 0.025 


1.041 0.002 

26.44 0.05 


30.00 0.03 

762.00 0.76 


0.010 0.001 

0.25 0.03 

These dimensions and tolerances apply to the material immediately after 

cutting and perforating. 

* In any group of four consecutive perforations, the maximum difference of 

pitch shall not exceed 0.001 inch and should be as much smaller as possible. 

** This dimension represents the length of any 100 consecutive perforation 



Proposed American Standard 
Cutting and Perforating Dimensions for 

32-mm Sound Motion Picture 

Negative and Positive Raw Stock 

December 1 948 

Page 2 of 2 Pages 


The dimensions given in this standard represent the practice of film manu- 
facturers in that the dimensions and tolerances are for film immediately after 
perforation. The punches and dies themselves are made to tolerances con- 
siderably smaller than those given, but owing to the fact that film is a plastic 
material, the dimensions of the slit and perforated film never agree exactly 
with the dimensions of the punches and dies. Shrinkage of the film, due to 
change in moisture content or loss of residual solvents, invariably results in 
a change in these dimensions during the life of the film. This change is 
generally uniform throughout the roll. 

The uniformity of perforation is one of the most important of the variables 
affecting steadiness of projection. 

Variations in pitch from roll to roll are of little significance compared to 
variations from one sprocket hole to the next. Actually, it is the maximum 
variation from one sprocket hole to the next within any small group that is 
important. This is one of the reasons for the method of specifying uniformity 
in dimension B. 

Thirty-two-millimeter release print stock is slit, after printing and develop- 
ing, to 16-mm. width. Since a possible error is involved in this slitting, the 
width of 32-mm. film is made 0.001" narrower than twice the width of 
standard 16-mm. film. This narrowing gives a tolerance of 0.001" in this 
secondary slitting operation. If the error of slitting exceeds this tolerance, 
one of the 16-mm. halves may exceed the width allowed for 16-mm. film 
and cause interference in the gate of a projector. In addition to errors of 
centering, there are errors caused by recurring variations in width. These 
errors will cause weave on the screen even though the maximum width of 
)he film may not be great enough to cause interference in the projector gate. 




Proposed American Standard 
Cutting and Perforating Dimensions for 

32-mm Silent Motion Picture 

Negative and Positive* Raw Stock 

December 1948 

Page 1 of 2 Pages 


^n n a 




a. ; 

n n I rr 


a a D a" 

d r 

L U E 

an a 





1.257 0.001 

31.93 0.025 


0.300 0.0005 

7.620 0.013 


0-.0720 O.OC04 

1.83 rh 0.01 


0.0500 0.0004 

1.27 0.01 


0.036 0.002 

0.91 0.05 


Not > 0.001 

Not > 0.025 


1.041 0.002 

26.44 0.05 


0.413 0.001 

10.490 0.025 


0.071 0.001 

1.803 0.025 


30.00 0.03 

762.00 0.76 


0.010 0.001 

0.25 0.03 

These dimensions and tolerances apply to the material immediately after 

cutting and perforating. 

* In any group of four consecutive perforations, the maximum difference of 

pitch shall not exceed 0.001 inch and should be as much smaller as possible. 

** This dimension represents the length of any 100 consecutive perforation 



Proposed American Standard 
Cutting and Perforating Dimensions for 

32-mm Silent Motion Picture 

Negative and Positive Raw Stock 

December 1941 

Page 2 of 2 Pages 


The dimensions given in this standard represent the practice of film manu- 
facturers in that the dimensions and tolerances are for film immediately after 
perforation. The punches and dies themselves are made to tolerances con- 
siderably smaller than those given, but owing to the fact that film is a plastic 
material, the dimensions of the slit and perforated film never agree exactly 
with the dimensions of the punches and dies. Shrinkage of the film, due to 
change in moisture content or loss of residual solvents, invariably results in 
a change in these dimensions during the life of the film. This change is 
generally uniform throughout the roll. 

The uniformity of perforation is one of the most important of the variables 
affecting steadiness of projection. 

Variations in pitch from roll to roll are of little significance compared to 
variations from one sprocket hole to the next. Actually, it is the maximum 
variation from one sprocket hole to the next within any small group that is 
important. This is one of the reasons for the method of specifying uniformity 
in dimension B. 

Thirty-two-millimeter release print stock is slit, after printing and develop- 
ing, to 16-mm. width. Since a possible error is involved in this slitting, the 
width of 32-mm. film is made 0.001" narrower than twice the width of 
standard 16-mm. film. This narrowing gives a tolerance of 0.001" in this 
secondary slitting operation. If the error of slitting exceeds this tolerance, 
one of the 16-mm. halves may exceed the width allowed for 16-mm. film 
and cause interference in the gate of a projector. In addition to errors of 
centering, there are errors caused by recurring variations in width. These 
errors will cause weave on the screen even though the maximum width of 
the film may not be great enough to cause interference in the projector gate. 




Proposed American Standard 
Cutting and Perforating Dimensions for 

32-mm on 35-mm Motion Picture 

Negative Raw Stock 

December 1948 

Page 1 of 2 Pages 











1.377 0.001 

34.98 0.025 


0.300 0.0005 

7.620 0.013 


0.0720 0.0004 

1.83 0.01 


0.0500 0.0004 

1.27 0.01 


0.096 0.002 

2.44 0.05 


Not > 0.001 

Not > 0.025 


1.041 0.002 

26.44 0.05 


30.00 0.03 

762.00 0.76 


0.010 0.001 

0.25 0.03 

These dimensions and tolerances apply to the material immediately after 
cutting and perforating. 

* In any group of four consecutive perforations, the maximum difference of 
pitch shall not exceed 0.001 inch and should be as much smaller as possible. 
' * This dimension represents the length of any 1 00 consecutive perforation 


Proposed American Standard 
Cutting and Perforating Dimensions for 

32-mm on 35-mm Motion Picture 

Negative Raw Stock 

December 1948 

Page 2 of 2 Pages 


The dimensions given in this standard represent the practice of film manu- 
facturers in that the dimensions and tolerances are for film immediately after 
perforation. The punches and dies themselves are made to tolerances con- 
siderably smaller than those given, but owing to the fact that film is a plastic 
material, the dimensions of the slit and perforated film never agree exactly 
with the dimensions of the punches and dies. Shrinkage of the film, due to 
change in moisture content or loss of residual solvents, invariably results in 
a change in these dimensions during the life of the film. This change is 
generally uniform throughout the roll. 

The uniformity of perforation is one of the most important of the variables 
affecting steadiness of projection. 

Variations in pitch from roll to roll are of little significance compared to 
variations from one sprocket hole to the next. Actually, it is the maximum 
variation from one sprocket hole to the next within any small group that is 
important. This is one of the reasons for the method of specifying uniformity 
in dimension B. 

This kind of 32 mm. film is made on 35 mm. stock so that it may be proc- 
essed on 35 mm. sprocketless negative developing machines. 

This film should not be made on nitrate base, because if this material 
were slit to 16 mm. it might be used on 'a projector with consequent danger 
of fire. 


JOSEPH H. McNABB, president and chairman of the Board of the 
Bell and Howell Company, died on January 5, 1949, in Chicago. 
Mr. McNabb was born in Canada, and received his technical edu- 
cation at the Collegiate Institute, St. Thomas, Ontario. He joined 
Bell and Howell in 1916, and in 1922, he was made chairman of the 

Mr. McNabb is well remembered for his activities in the profes- 
sional and amateur motion picture fields, as an organizer and execu- 
tive, and for his contributions to the war effort, particularly in the 
field of optics. He was a Fellow of the Society of Motion Picture 

65th Semiannual Convention 

Hotel Statler, New York, N. Y., April 4-8, i949 

The 65th Semiannual Convention of the Society will be held at the Hotel Static 
(formerly Hotel Pennsylvania) in New York City from April 4 to 8, 1949, inclu 
sive. Authors should submit complete manuscripts of their papers to Miss Helen 
Stote, SMPE JOURNAL Editor, Room 912, 342 Madison Ave., New York 17, N. Y., 
by March 1 . All abstracts should be sent to Mr. Seeley before February 15. The 
Chairman and Vice-Chairmen of the Papers Committee are as follows: 

N. L. Simmons, Chairman 
6706 Santa Monica Blvd. 
Hollywood 38, California 

Joseph E. Aiken, Vice-Chairman R. T. Van Niman, Vice-Chairman 

116 N. Galveston 4331 West Lake Street 

Arlington, Va. Chicago 24, Illinois 

Lorin Grignon, Vice-Chairman H. S. Walker, Vice-Chair man 

20th Century-Fox Films Corp. 1620 Notre Dame Street, W. 

Beverly Hills, California Montreal, Que., Canada 

E. S. Seeley, Vice-Chair man 
Altec Service Corp. 
161 Sixth Avenue 
New York 13, N. Y. 


The Monday afternoon session will be a forum on the many aspects of the use 
of films in television, including statements by several qualified experts, followed 
by a question-and-answer period. There will be a demonstration Monday night. 

Tuesday will be devoted to technical and economic aspects of television. 

On Wednesday, April 6, there will be held a Symposium on High-Speed Pho- 
tography. Authors and delegates from several foreign countries will attend this 
second SMPE Symposium on this subject. 

Thursday and Friday will be devoted to papers dealing with Motion Picture 
Studio and Theater Equipment, Recording, Films, and Processing. 


The usual get-together Luncheon will be held in the Georgian Room at 12:30 
P.M., Monday, April 4. 

The Cocktail Hour and Banquet will be in the Georgian Room on Wednesday 
evening, April 6. 



Monday, April 4, 1949 


Conference Room 9, 

18th floor 



Georgian Room 
Salle Moderne 

Salle Moderne 

Tuesday, April 5, 1949 


Conference Room 9, 

18th floor 


Salle Moderne 

Salle Moderne 


Wednesday, April 6, 1949 

Conference Room 9, 

18th floor 


Salle Moderne 

Salle Moderne 

Georgian Room Foyer 
Georgian Room 

Thursday, April 7, 1949 

Salle Moderne 

Salle Moderne 

Fiday, April 8, 1949 

Salle Moderne 

Salle Moderne 


Mrs. Earl I. Sponable will serve as Hostess, and Mrs. W. H. Rivers will assist 
her. Ladies' reception and registration headquarters will be located hi Room 129. 


Room-reservation cards have been mailed to the membership. These should be 
filled in and returned promptly to the hotel. 



Motion Picture Test Films 

rpiHE TEST-FILM PROGRAMS of the Society of Motion Picture Engineers and the 
J- Motion Picture Research Council have been combined in recent years in an 
attempt to avoid unnecessary duplication of sources and catalogs that had been ') 
confusing to those not familiar with both organizations. The result has been im- 
proved service to test-film users in the form of better integrated programs for *; 
developing new films and a single catalog. In 1947 the first catalog, listing the | 
twenty-nine films then available, was published as part of the August issue of the 
JOURNAL and was later reprinted by most motion picture trade magazines. It is \ 
now out of print but a revised version will be off the press shortly with copies j 
available to all film users. 

All of the test films, both 35-mm and 16-mm, are now supplied on safety stock. 
They are furnished as a service to the industry and because neither organization 
has provision for extending credit, all test films are sold on a "cash-with-order" 
basis at the prices shown here. Orders received without payment enclosed will be 
shipped C.O.D. 

Prices shown do, however, include shipping charges to all points within the 
United States. Where required by purchasers in other countries', shipping charges 
will be billed to the customer. If necessary, pro forma invoices will be supplied 

When placing an order with either the Society of Motion Picture Engineers, 
Inc., 342 Madison Avenue, New York 17, N. Y., or the Motion Picture Research 
Council, Inc., 1421 North Western Avenue, Hollywood 27, California, be sure to 
specify clearly 

1. Name of the film 

2. Code Number 

3. Price 

4. Enclose payment with order. 


Code Length 

Test Film No. (in Feet) Price 

35-Mm Visual Test Film VTF-1 450 $25.00 

Focus-and-Alignment Section VTF-FAS 100 8.00 

Travel-Ghost Target Section VTF-TGS 100 8.00 

Jump-and-Weave Target Section VTF-JWS 100 8.00 

35-Mm Theater Sound Test Film ASTR-3 500 17.50 

35-Mm Multifrequency Test Film 

Ttfpe A Laboratory Type APFA-1 450 25.00 

Type B Service Type ASFA-1 300 .17.50 

35-Mm Transmission Test Film TA-1 250 17.50 

35-Mm Buzz-Track Test Film ABZT-1 50 min* 0.04/ft 

Motion Picture Test Films 

35-Mm Scanning-Beam Illumination 
Test Film 

Type A 17-Position Track 
Type B Snake Track 





35-Mm Sound-Focusing Test Film 
Type A 9000-Cycle Track 
Type B 7000-Cycle Track (Area) 
Type C 7000-Cycle Track (Den- 



50 min 
50 min 

50 min 



35-Mm 3000-Cycle Flutter Test Film 


50 min 


35-Mm 1000-Cycle Balancing Test 

For Two Machines 




For Three Machines 




1000-Cycle Test Film 


50 min 


35-Mm Multifrequency Warble Test 




16-Mm Sound-Projector Test Film 




16-Mm Multifrequency Test Film 




16-Mm Buzz-Track Test Film 




16-Mm Scanning-Beam Illumination 
Test Film 

Laboratory Type 
Service Type 





16-Mm Sound-Focusing Test Film 
Laboratory Type 
Service Type 

Z22. 42-7000 
Z 22. 42-5000 



16-Mm 3000-Cycle Flutter Test Film Z22.43 100 27.50 

16-Mm 400-Cycle Signal-Level Test 
Film Z22.45 100 27.50 

* Minimum. 


Book Review 

An Introduction to Color, by Ralph M. Evans 

Published (1948) by John Wiley and Sons, Inc., 440 Fourth Ave., New York 16, 
N. Y. 324 pages + X pages + 3-page general bibliography + 12-page index. 
269 illustrations + 15 color plates. 7V2 X 9 3 A inches. Price, $6.00. 

Color may be defined as a hue of the rainbow or spectrum, or as a tint produced 
by a mixture of such hues, or as a paint or a pigment, or in various other ways for 
ordinary usage. To the scientist, however, color has a more specific meaning de- 
pending upon its use and application. 

To the physicist color is one of the characteristics of the radiant energy known 
as light. He is concerned with the properties of light in terms of wavelengths 
or frequencies, and with its intensity, both absolute and comparative. He has 
devised many instruments, such as monochrometers, spectrophotometers, radi- 
ometers, photoelectric cells, and the like, and employs many different light 
sources for the emission of light, all for the purpose of describing and defining 
light and color in terms of his accepted constants such as lumens, foot-candles, 
ergs, joules, and so forth. 

The psychophysicist, on the other hand, takes the "color" of the physicist and 
relates it to the sensitivity of the human eye. In so doing he evaluates the 
physicist's "color" in terms of an average observer usually by the use of mechan- 
ical means. He is interested in the entire system composed of the light source, 
the radiant energy, and the human eye. 

The psychologist studies the action of the mind in perceiving and interpreting 
light and color, and explains, for example, why the same physical light is not 
always seen as the same color. 

Some time ago the Committee on Colorimetry of the Optical Society of America, 
recognizing the complexities and interrelationship of the many variables involved 
in the field of color, tabulated them in chart form under the head of physics, 
psychophysics, and psychology. For the first time, however, to this reviewer's 
knowledge, a book written by an expert in the field has been published in order 
that the reader interested in color may understand, correlate, and interrelate all of 
these many variables. 

It would be difficult to find a man better suited for the task than Mr. Evans. 
An authority in the field of color, he has for many years been conducting research 
on visual effects in photography, and has delivered many lectures, and written 
numerous papers on the sub j ect of color. His first book on the sub j ect is a milestone. 

In "An Introduction to Color" Mr. Evans covers the physics of color in the 
first six chapters with headings as follows: Chapter I, "Color and Light"; Chap- 
ter II, "The Physical Nature of Light"; Chapter III, "Light Sources"; Chapter 
IV, "Illumination"; Chapter V, "Colored Objects"; Chapter VI, "The Physics 
of Everyday Color." Without delving too deeply into the subject, the reader is 
given a brief but adequate grasp of the basic physics of light and color. 

Book Review 

The psychophysics of light and color are discussed in Chapter VII, entitled, 
"Color Vision," including a brief review of the elementary physiology of the eye 
and the functions of the various brain centers. Mr. Evans discusses light and 
color as evaluated by the eye as a fixed sensitivity receptor. 

Having treated the physical and psychophysical aspects of light and color in 
earlier chapters, Mr. Evans then directs his attention to the psychological aspects 
of vision. In Chapters VIII through XI, entitled, Chapter VIII, "The Visual 
Variables of Color"; Chapter IX, "Perception and Illusion"; Chapter X, 
"Brightness and Perception"; Chapter XI, "Color Perception." Mr. Evans 
reveals the part which the mind plays as the final interpreter of light and color 
phenomena. He makes clear why the same physical light can appear different to 
the eye under different conditions and in different environments. Illusions and 
other unexpected phenomena which exist in the form and shape of objects can 
also exist in light and color, and some of the bases of their interrelationship are 
shown in these chapters. 

In the subsequent eleven chapters of the book Mr. Evans warms up to problems 
which are more purely those of color such as the measurement of color, the speci- 
fications of color, a thorough review of the various color systems, their assumptions, 
tools and methods of use and application, color differences and color names, 
additive and subtractive mixtures of colors, and transparent and nontransparent 
color mixtures (paints). The author uses the same approach in discussing these 
problems as he did in considering the relationship of the physics, the psycho- 
physics, and the psychology of light and color. 

In the last three chapters of the book the role of color in photography, art, de- 
sign, and abstraction is studied and interpreted in the light of the previous chapters. 

The book is extensively illustrated, and noteworthy are the clear explanations 
and titles of the many graphs and pictures. "An Introduction to Color" treats 
difficult subjects with clarity and simplicity. Definitions and usages of words 
unique to their field of science are adequately explained. 

The author's emphasis through explanation and re-explanation, even if some- 
times repetitious, will be helpful and pleasing to the many teachers and students 
who will unquestionably use this book for study and reference purposes. 

The absence of much of the mathematical background of the many concepts 
discussed will be of benefit to the relatively untrained reader interested in color, 
and will not prove a serious deficiency for the trained scientist. 

The book is highly recommended by this reviewer for the serious worker in any 
field in which color is an important factor, and is a "must" for the artist, tech- 
nician, and scientist in color photography and related arts. 


Technicolor Motion Picture Corporation 

Hollywood, Calif. 


1949 Nominations 

THE 1939 NOMINATING COMMITTEE, as appointed by the President of the 
Society was confirmed "by the Board of Governors at its January meeting. 

D. E. HYNDMAN, Chairman 

Room 626 342 Madison Avenue 

New York 17, N. Y. 


General Precision Equipment Corp. Technicolor Motion Picture Corp. 

63 Bedford Road 6311 Romaine Street 

Pleasantville, N. Y. Los Angeles 28, Calif. 

Research Laboratories A - N ' GOLDSMITH 

National Carbon Company 597 Flfth Avenue 

Box 6087 New York 17 > N ' Y- 

Cleveland 1, Ohio T T GOLDSMITH 

F. E. CAHILL, JR. A H en B . Du Mont Laboratories 

Warner Bros. Pictures, Inc. 2 Main Avenue 

321 W. 44 Street Passaic N J 
New York 20, N. Y. 


General Electric Company Western Electric Company 

Nela Park 6601 Romaine Street 

Cleveland 12, Ohio Hollywood 38, Calif. 

All voting members of the Society who wish to submit recommendations for 
candidates to be considered by the Committee as possible nominees, are requested 
to correspond directly with the Chairman or any of the members of the Nominat- 
ing Committee. Active, Fellow, or Honorary Members are authorized to make 
these suggestions which must be in the hands of the Committee by May 1, 1949. 

There will be eight vacancies on the Board of Governors as of January 1, 1950, 
which must be filled. Those members whose terms of office expire are: 

Financial Vice-President D. B. JOY 

Engineering Vice-President J. A. MAURER 

Treasurer R. B. AUSTRIAN 

Governor. . . .A. W. COOK Governor. . . .P. J. LARSEN 

Governor. . . .JAMES FRANK, JR. Governor. . . .G. E. SAWYER 

Governor. . . .L. T. GOLDSMITH 

The recommendations of the Nominating Committee will be submitted to the 
Board of Governors for approval at the July meeting. The ballots will then be 
prepared and mailed to the voting members of the Society forty days prior to the 
Annual Meeting of the Society. This is the business session held during the Fall 
Convention, which this year will be in Hollywood, California, October 10-14. 

D. E. HYNDMAN, Chairman, 
Nominating Committee 


Optical Society Meeting 

rr^HE OPTICAL SOCIETY OP AMERICA will hold its Winter Meeting in New York 
1 City, on March 10, 11, and 12, 1949, with headquarters at the Hotel Statler. 

A special feature of the meeting will be a symposium on luminescence, to be 
held on the first day of the meeting under the guidance of G. R. Fonda of the Gen- 
eral Electric Company. This symposium will be high-lighted by "A Survey of 
Present Methods Used to Determine the Optical Properties of Phosphors," by 
Wayne B. Nottingham, Massachusetts Institute of Technology, and "Review of 
the Interpretations of Luminescence Phenomena," by Fred E. Williams, Research 
Laboratory, General Electric Company. 

Anyone desiring to attend this meeting should notify Dr. G. R. Fonda, General 
Electric Company, Schenectady, N. Y., at least three weeks in advance of the 

The sessions for contributed papers on Friday and Saturday will be preceded by 
the following invited papers: "The Ruling of Large Diffraction Gratings," by 
G. R. Harrison, Massachusetts Institute of Technology; "Measurements of Size 
and Shape of Large Molecules by Light Scattering," by P. Debye, Cornell Uni- 
versity; and "A Color Translating Ultraviolet Microscope," by E. H. Land, Polar- 
oid Corporation. 

This meeting will be open to nonmembers of the Society, and all interested 
persons are cordially invited to attend. 

Current Literature 

THE EDITORS present for convenient reference a list of articles dealing with 
subjects cognate to motion picture engineering published in a number of se- 
lected journals. Photostatic or microfilm copies of articles in magazines that are 
available may be obtained from The Library of Congress, Washington, D. C., or 
from the New York Public Library, New York, N. Y., at prevailing rates. 

American Cinematographer International Photographer 

29, 12, December, 1948 20, 12, December, 1948 

Latensification (p. 409) H. W. New "Spectra" Meter (p. 7) 

Mobile Camera Lab. (p. 410) W. M. 

CONE International Projectionist 

Karl Freund Introduces the Con- 23, 12, December, 1948 

tract Lineup System (p. 413) The New British SUPA Projector 

R. LAWTON (p. 21) H. HILL 
Audio Engineering 

33, 1, January, 1949 Radio and Television News 

Psycho-Acoustic Aspects of Higher 40, 6, December, 1948 

Quality Reproduction (p. 9) C. J. The Recording and Reproduction of 

LEBEL Sound. Pt. 22 (p. 48) O. READ 


New Products 

T^urther information concerning the material described below can 
JL be obtained by writing direct to the manufacturers. As in the case 
of technical papers, publication of these news items does not consti- 
tute endorsement of the manufacturer's statements nor of his products. 

Brenkert Film Projector 

A new 35-mm Brenkert film projec- 
tor, especially designed for the medium 
size theater which must operate on a 
conservative budget, was introduced 
recently by the RCA Theater Equip- 
ment Section, 36 W. 49 St., New York 
20, New York. 

Engineering and performance fea- 
tures of the new model include a design 
for the rear shutter blade which sup- 
plies good ventilation to the projection 

aperture for cooling purposes, and an 
operating compartment that is oilfree 
and roomy, providing maximum space 
for threading the projector. 

An important feature of the BX-60 is 
the automatic lubrication system. All 
rotating shafts running through the 
main case casting are equipped with oil 
baffles, so that shaft bearings are con- 
tinuously lubricated throughout their 
length, but no oil can leak into the 
operating compartment. 

A large door on the operating side of 
the projector exposes the entire film 
compartment and two glass-covered 
openings permit the operator to ob- 
serve the film loops above and below 
the film trap while the mechanism is in 
operation. Quick access to the shutter 
blades and the rear of the film trap is 
gained by removal of a panel on the 
operating side which is held in place 
by two thumbscrews. A filter glass is 
provided in this panel for viewing the 
light on the aperture. 

The intermittent mechanism in the 
BX-60 is identical to that in the larger 
Brenkert BX-80 projector. Unit con- 
struction is used to facilitate easy, 
quick, and accurate servicing. All 
units are doweled to the main frame 
for correct alignment of parts, thereby 
maintaining the accuracy built into the 

Journal of the 

Society of Motion Picture Engineers 



Theater Television* 243 

Research Council Small Camera Crane ANDRE CROT 273 

Experiment in Stereophonic Sound LORIN D. GRIGNON 280 

Single-Element Unidirectional Microphone 


16-Mm Film Phonograph for Professional Use . . CARL E. HITTLE 303 
Frequency-Modulated Audio-Frequency Oscillator for 

Calibrating Flutter-Measuring Equipment 


Silent Playback and Public-Address System 


New Automatic Sound Slidefilm System W. A. PALMER 320 

Magnetic Device for Cuing Film JAMES A. LARSEN 326 

Improved 35-Mm Synchronous Counter 

Proposed Standards for 16-Mm and 8-Mm Picture Apertures. . 337 

65th Semiannual Convention 349 

Charles G. Weber 353 

Section Meeting 354 

Armed Forces Communications Association 354 

1949 Nominations 356 

Book Reviews: 

"Sound and Documentary Film," by K. Cameron (Foreword 
by Cavalcanti) 

Published (1947) by Sir I. Pitman and Sons, Ltd. 

Reviewed by Glenn E. Matthews 357 

^Discharge Lamps," by H. K. Bourne 

Published (1948) by Chapman and Hall, Ltd. 

Reviewed by F. E. Carlson 357 

Standards Recommendation 358 

New Products 359 


Chairman Editor Chairman 

Board of Editors Papers Committee 

Subscription to nonmembers, $10,00 per annum; to membera, $6.25 per annum, included in 
their annual membership dues; single copies, $1.25. Order from the Society's General Office. 
A discount of ten per cent is allowed to accredited agencies on orders for subscriptions and 
single copies. Published monthly at Easton, Pa., by the Society of Motion PictureEngineers 
Inc. Publication Office, 20th & Northampton Sts., Easton, Pa. General and Editorial Office, 
342 Madison Ave., New York 17, N. Y. Entered as second-class matter January 15, 1930, 
at the Post Office at Easton, Pa., under the Act of March 3, 1879. 

Copyright, 1949, by the Society of Motion Picture Engineers, Inc. Permission to republish 
material from the JOURNAL must be obtained in writing from the General Office of the Society. 
Copyright under International Copyright Convention and Pan-American Convention. The 
Society is not responsible for statements of authors or contributors. 

Society of 

Motion Picture Engineers 

342 MADISON AVENUE NEW YORK 17, N. Y.--TEL. Mu 2-2185 


Earl I. Sponable 
460 W. 54 St. 
New York 19, N. Y. 

Peter Mole 

941 N. Sycamore Ave. 
Hollywood 38, Calif 


Loren L. Ryder 
5451 Marathon St. 
Hollywood 38, Calif. 

Clyde R. Keith 
120 Broadway 
New York 5, N Y 

Robert M. Corbin 
343 State St. 
Rochester 4, N. Y. 

William C. Kunzmann 
Box 6087 
Cleveland 1, Ohio 



VICE-PRESIDENT Ra i ph 3. Austrian 

John A. Maurer 1270 Avenue of The 

37-0131 St. Americas 

Long Island City 1, N. Y. New York 20, N. Y. 


David B. Joy 
30 E. 42 St. 
New York 17, N. Y. 



Alan W. Cook 
25 Thorpe St. 
Binghamton, N. Y. 

Lloyd T. Goldsmith 
Warner Brothers 
Burbank, Calif. 

Paul J. Larsen 
508 S. Tulane St 
Albuquerque, N. M. 

Gordon E. Sawyer 
857 N. Martel Ave. 
Hollywood 46, Calif. 


James Frank, Jr. 
426 Luckie St., N. W. 
Atlanta, Ga. 

William H. Rivers 
342 Madison Ave. 
New York 17, N. Y. 

Sidney P. Solow 
959 Seward St. 
Hollywood 38, Calif. 

R. T. Van Niman 
4431 W. Lake St. 
Chicago 24, 111. 


Herbert Barnett 
Manville Lane 
Pleasantville, N. Y. 

Fred T. Bowditch 
Box 6087 
Cleveland 1, Ohio 

Kenneth F. Morgan 
6601 Romaine St. 
Los Angeles 38, Calif. 

Norwood L. Simmons 
6706 Santa Monica Blvd. 
Hollywood 38, Calif. 

Theater Television 


DURING THE YEARS 1944, 1945, and 1946, the Theater Television 
Committee of the SMPE worked on many of the various engi- 
neering problems related to placing television in the motion picture 
theater. Considerable time and effort were expended in collecting in- 
formation on such subjects as existing theater architecture, available 
projection equipment, picture quality to be expected, and many other 
problems which would be of interest to the motion picture industry if 
it intended to use this new medium. Also during these years, Paul 
J. Larsen, together with other representatives of the Society, ap- 
peared before the Federal Communications Commission and suc- 
ceeded in obtaining for the motion picture industry frequency alloca- 
tions for theater-television use on an experimental basis only. 

The Motion Picture Association was then approached with re- 
peated requests that it co-operate in the television work of the So- 
ciety if it had any reason to believe that theater television would be 
practical. Neither producers nor distributors, 'however, were in- 
terested in theater television at that time nor were they particularly 
concerned about television as a competitive entertainment medium. 
The exhibitors on the other hand, showed some concern but did not 
wish to take any active measures either on their own or with the 
Society. The general attitude seemed to be that it might be possible 
to buy into the television industry at some future date and thereby 
save the high cost of research and development. 

In November, 1946, a point had been reached in the technical work 
where it was believed a definite statement of interest by the motion 
picture industry was required if the work were to continue. In addi- 
tion, public hearings before the FCC were scheduled for early 1947 
at which it was proposed to reallocate to other services the frequencies 
formerly provided for experimental theater use. 

In spite of the lack of interest shown, the Society again undertook 
having a brief prepared and Mr. Larsen, then chairman of the Theater 
Television Committee, appeared before the FCC on February 4, 1947. 
(See FCC Docket No. 6651.) Immediately preceding the hearings a 
telegram was received from Eric Johnston, president of the Motion 
Picture Association, endorsing the SMPE's stand. A similar telegram 
was sent to the FCC three weeks following the hearings by Donald 


Xelson, then president of the Society of Independent Motion Picture 

The decision of the FCC was handed down early in 1948 and while 
it did not provide specific frequency allocations for theater use, it did 
make available certain frequencies which could still be used by the 
motion picture industry for experimental purposes. 

Consequently, even though the motion picture industry's position 
is weaker from the standpoint of obtaining a permanent part of the 
radio-frequency spectrum, the Theater Television Committee de- 
cided to continue its engineering work. 

The members agreed to draw up as comprehensive a report as 
possible outlining the present state of the art in so far as it regards the 
motion picture theater and again to seek the co-operation of the in- 
dustry as a whole. Such a report is contained in the following pages. 

In this report an attempt has been made to present the information 
in as nontechnical language as possible so that those not closely asso- 
ciated with the television industry will be aware of the tremendous 
strides made in this new medium during the last few years. 

Government regulations involved, types of theater equipment 
available, and distribution facilities now in existence are described in 
a manner which it is hoped will be of value to all branches of the 
motion picture industry. 


The following is a general statement as to the regulations and pro- 
cedures which in part govern the establishment of a theater-tele- 
vision system of urban, intercity, or national scope. It is intended 
only as a general and partial guide and must be supplemented by de- 
tailed information prior to any definite action on the part of those 
planning theater-television-program syndication either locally or on a 
nation-wide scale. Detailed information on the rules and regulations 
of the Federal Communications Commission can be secured by ad- 
dressing the Secretary of the Commission in Washington. 

No Federal license or governmental permission is required for the 
establishment of a theater-television receiving station, either for the 
reception of programs by wire or coaxial cable or by radio. However, 
municipal regulations may control the placement of masts or other 
structures on roofs, the guying of reinforcements of such structures, 
and the safety of any electrical wiring of permanent nature installed 
in the theater. If high towers are erected for reception ; and if these 


are so located that they may become an obstacle or hazard to aerial 
navigation, it is possible that the Civil Aeronautics Authority must 
grant approval prior to the erection of such facilities. 

If one or more theaters in a given city are to receive a television 
program from central studios, means must be provided for carrying 
the program from the central studios to the theaters in question. The 
program may originate from live-talent presentations in the studio 
(or at remote pickup points such as sports arenas, legitimate theaters, 
or the like), or they may originate from film records previously made. 
They may be carried to the theaters by means of specially equalized 
telephone lines or by coaxial cables, either of which presumably will 
be furnished by the local telephone company or other public-utility 
common carrier in the communications service. Alternatively, the 
programs may be sent by narrow radio beams from a central trans- 
mitting station to the individual theaters where they are received on 
highly directional antennas or aerials. 

If radio beams are to be used, it becomes necessary to secure the 
approval of the Federal Communications Commission and to receive a 
construction permit and, thereafter, a station license to permit the 
operation of the transmitter which sends the studio program to the 
various theaters. 

The transmissions in question are not broadcast (that is, addressed 
to the general public), and a broadcasting license would not be re- 
quired from the FCC. The transmissions are rather of the type 
known as multiple-addressee messages which are private communica- 
tions addressed by a single sender to a group of recipients, each of 
whom receives the same message. Such messages, unlike broadcasts, 
are private in nature, and are not legally available to the public. 

Any central transmitter erected to send television programs to a 
group of theaters will require a suitable tower to support the trans- 
mitting equipment. If this tower is a potential aeronautical hazard, 
authorization will be required from the Civil Aeronautics Authority 
for its erection. Municipal codes relative to the establishment of 
towers in residential districts must also be considered. 

The FCC has taken the present stand that the distribution of ma- 
terial of a private nature, such as is here contemplated, falls within 
the scope of the common carriers and that it should, therefore, be 
handled by a telephone or telegraph company. It is not known 
whether the FCC will permanently adhere to this policy, particularly 
in the case of urban television studios and transmitters for syndication 


of programs to theaters. In any case, the channels (frequencies) 
assigned to such television transmissions necessarily would be different 
from those used for television broadcasting. The channels would 
probably be at considerably higher frequencies, would be wider (to 
accommodate possibly higher-detail pictures, or color pictures, or 
both), and might differ from the broadcasting channels in other re- 
spects as well. 

The SMPE has previously requested allocations of channels from 
the FCC for commercial theater television, but that request has not 
as yet been granted. It is not known whether the Commission ulti- 
mately would grant such channels but it is believed that their grant 
would require the following steps : 

An individual theater owner planning to establish a television 
service to its theaters would first apply for an experimental license 
from the FCC to permit him to transmit his programs, purely experi- 
mentally and noncommercially, for a specific period. He would be 
obligated to describe his plans clearly, and to report from time to time 
to the Commission on his technical progress. 

If his experiments were successful, he might then ask that his ex- 
perimental license be converted into a commercial license permitting 
normal and continued operation during the period of the license 
(which might be set at three years, or some similar period). The 
Commission then doubtless would hold hearings to determine the need 
for and desirability of the service in question. If it found that the 
service was useful and necessary and that channels were available, it 
would then grant the corresponding commercial license. City-alloca- 
tion hearings would also be held. 

It should be added that each theater chain in the same city would 
require its own transmitting facilities or wire network (unless an in- 
terchange of programs or the common use of a single transmitter were 
acceptable to all involved). That? is, for completely independent 
service each theater group would require its own transmitter, its own 
receivers, and its own channel allocations from the FCC. An individ- 
ual theater owner or theater circuit might arrange for transmission 
of its programs by an existing television station or licensee in a manner 
similar to current commercial-television broadcasts. 

In addition, for remote pickups outside the main studios, it becomes 
necessary (if radio is to be used to carry the program from the remote 
point to the central studios for retransmission to the theaters) to 
secure an FCC license for the radio-beam transmitter which will 


carry such programs. Channels are presently available for that 
purpose and might, for good reason, be secured from the FCC. 

When nation-wide syndication is involved, it becomes necessary to 
interconnect the central studios of each network in the cities in which 
it serves theaters by means of coaxial cables (of the telephone com- 
pany) or by radio relay (either supplied by the telephone company or 
other common carrier, or else established by special permission of the 
FCC as the result of a change in its present policy). Such radio-relay 
systems consist of a number of repeater stations about thirty miles 
apart, each of which receives the program from the preceding station 
and automatically carries it forward to the next station. These relay 
or "booster" stations may be unattended and subject only to occa- 
sional inspection and the replacement of expendable material. 

In brief, those considering the use of radio in theater television are 
particularly directed to the following basic points which may be novel 
to those not familiar with the field of radio communication. 

1. Unlike the motion picture field, television by radio is subject to 
numerous governmental regulations and controls. Consequently, 
those entering theater television, and using radio transmission, must 
be thoroughly familiar with governmental rules and procedures and 
be governed thereby for their own protection. 

2. In the second place, television by radio requires so-called wide- 
channel assignments by the FCC. Such channels are scarce and 
much sought. Accordingly, nonuse of such channels amost inevitably 
leads to their pre-emption by others. 

3. Accordingly, if theater television is to secure such radio chan- 
nels, it must promptly request their assignment. However, a mere 
request is not generally sufficient to persuade the FCC to grant 
channels. Usually financial responsibility, definiteness of construc- 
tion and operating plans, niture of ownership and affiliation, willing- 
ness to report all technical (and perhaps program) progress, and other 
obligations must be made sufficiently clear and definite to the Com- 
mission to justify the assignment of channels. There seems little 
likelihood that a vague expression of general interest or intent will 
lead to channel assignments. 

4. In any case, even an otherwise satisfactory application for 
channels must be denied if no available and noninterfering channels 
are any longer existent, because of prior assignments. The conclu- 
sions to be drawn are evident. 



Two basic systems of large-screen theater television are currently 
being evaluated in this country. One is the instantaneous or direct- 
projection system by which high-brilliance cathode-ray-tube images 
are projected by means of an efficient reflective optical system; the 
other is the storage or intermediate-film system using standard motion 
picture projection technique, after television images have been photo- 
graphed or transcribed on motion picture film and suitably processed. 

Although neither type is commercially available in production 
quantities, the rapid progress of the art warrants description of equip- 
ment in experimental installations. 


The direct-projection system consists of three major optical ele- 
ments: 1. the projection cathode-ray tube which is the source of the 
light image; 2. the optical system which projects the image into the 
screen; and 3. the screen from which the final image is viewed. In 
addition to the optical elements of the system, which are housed in a 
projection barrel, the electronic auxiliaries include a control console 
containing the associated television equipment, and a power-supply 
rack. A typical experimental arrangement of this equipment as in- 
stalled in a theater is shown in Fig. 1. 

Projection Cathode-Ray Tube The cathode-ray tube used in the 
direct-projection system is similar to the direct- viewing tube used in 
the conventional television receiver, except that projection tubes 
have a much greater light output resulting from higher voltage opera- 
tion for which they are designed. 

Since available television for commercial operation is not adapted, 
at the present state of the art, to the use of supplementary light 
sources as are motion pictures, the brightness of the image available 
for projection depends upon the efficiency of the cathode-ray tube and 
the operating potentials. An average picture on a projection tube 
will draw a beam current of approximately 1 milliampere at a poten- 
tial of 80 kilovolts. This is a power of only 80 watts. With a screen 
efficiency of 5 candle power per watt this represents a light output of 
400 candle power. 

Optical System Typical optical systems employed in large-screen 
television have been discussed in various issues of the SMPE JOURNAL, 
and will therefore be only briefly reviewed here. The familiar refrac- 




tive projection optics used in motion picture film projectors deliver 
approximately 6 per cent of the light from the arc-light source to the 
screen. On the other hand, the reflective optics developed for tele- 
vision deliver 30 per cent of the light output from the cathode-ray 
tube to the screen. 

Reflective optics have been designed for large-screen projection of 
pictures up to 18 X 24 feet. One system for a 7'/2 X 10-foot screen 
uses a 21-inch mirror, a 14-inch "lens" (correction plate), and a 7-inch, 
50-kilovolt, cathode-ray tube. The largest system built so far con- 

Fig. 1 Typical direct-projection system. 

sisted of a 42-inch mirror, a 26-inch "lens," and 12- and 15-inch pro- 
jection tubes operating at 80 kilovolts. The throw was fixed at 40 feet 
and by changing the cathode-ray tube either a 15- X 20- or an 18- X 
24-foot picture was shown. The magnification is fixed by the mirror 
radius. High present production cost of large-mirror systems seem- 
ingly indicate the advisability of concentrating on smaller optics and 
increasing the voltage capabilities of smaller cathode-ray tubes (7- 
inch) in order to make a compromise system which might be successful 


Viewing Screen The viewing screen forms the third and final optical 
element of direct-projection television. Standard motion picture 
screens have a diffuse surface which distributes the light more or less 
uniformly in all directions. Since the distribution is nondirectional a 
great deal of light is lost to the ceiling and floor. Directivity, if it 
could be obtained in the vertical plane, would concentrate the light 
where it would be most useful and effect an important increase in 
efficiency. Beaded screens have been made to control the direction of 
the reflected light from the screen, but the directivity pattern, while 
showing a gain of 2, restricts the horizontal reflective pattern and 
tends to reflect a great deal of illumination back into the optical sys- 
tem where it reduces the contrast of the projected image. Develop- 
ments in directional screens now underway promise gains as high as 3. 
A lenticular screen of this type was successfully used in the Fox 
Theater in Philadelphia where a 15- X 20-foot picture was shown 
featuring the 1948 Louis- Walcott fight. This screen is embossed on 
an aluminized surface, with small convex-lens elements to control the 
directivity pattern. The observed results were excellent and a gain of 
two and one-half times was measured. 

Such is not the case with a normal translucent screen (rear projec- 
tion) since the light comes from a relatively small source and is a 
diverging cone of light at the screen. (The usual translucent screen 
receives direct rays which are normal to the center of the screen but 
diverge nearer the edges resulting in a bright spot in the center of the 
screen.) A field lens can be used on the rear of the translucent screen 
to direct the rays in a parallel pattern, and hence give more uniform 
illumination over the entire screen, or by a modification the pattern 
may be made to suit almost any application. Such a field lens may be 
applied only in small screens as in the home type of projection re- 
ceiver where a molded-plastic screen can be used. A compromise 
screen of high-density translucent material can be made, but the gain 
will be low and the directivity pattern becomes very sharp. 

Equipment Elements The current design trend for direct-projection, 
systems is to break the equipment into several discrete units : 1 . The, 
optical housing containing the mirror, lens, cathode-ray tube and itsj 
associated deflection coil, and a cooling system for the cathode-ray; 
tube; 2. The control console containing the critical television ele-j 
ments such as the video amplifier and deflection circuits as well as the; 


operating control panel; 3. The auxiliary power equipment consisting 
of a power-supply rack and a high-voltage power unit. 

Equipment Location Various locations have been suggested and 
tried for this type of projection-television equipment. The present 
throw limitation makes the normal booth installation impracticable. 
Longer throw systems up to 65 feet, can be made but again the cost 
and size factor rule them out. Rear projection might seem ideal for 
short-throw systems, but the screen directivity is too sharp to make 
this practicable. If it were economical to waste a great deal of light 
on a very dense screen there might be some compromise possible in 
this direction. Another important consideration in selecting the 
location is the projection angle because the limited depth of focus of 
the short optical system demands operation with the screen normal to 
the projection axis. 

The installation requirements are peculiar to the optical system 
employed and ideally would locate the optical housing on the front of 
the balcony. Alternatively in a nonbalcony house, the optical housing 
may be located either on a special ceiling suspension or in the orches- 
tra. The control console should not be more than 15 feet from the 
optical housing because of circuit requirements, which usually dictates 
its placement at the balcony rail. The balance of the equipment can 
be remotely placed at any convenient point, but cost will probably 
indicate a location less than 100 feet from the optical housing. 

Performance Picture quality from large-screen television projectors 
is now limited by the quality of the transmitted signals. The capa- 
bilities of the projection system are equal to the best studio television 
equipment and any deterioration of the signal between the camera and 
the projector causes an inferior picture on the screen. Experience has 
shown that with a picture of suitable quality it is possible to produce 
results acceptable to critical audiences. Present transmission of tele- 
vision pictures on standard channels is limited in bandwidth so that the 
proj ected pictures actually have about 300 lines resolution. If the pic- 
tures were transmitted by microwave relay, the entire capability of the 
projection system of approximately 350 to 450 lines could be utilized. 
A television system specifically designed for theater use will no 
doubt be a private system using ultrahigh-frequency channels and all 
of the equipment and techniques of operation will be improved to 
utilize the present standards to the fullest extent. 





Two basic image storage television projection systems are being in- 
vestigated. The first uses motion picture film as the intermediate 
storage medium while the second employs electronic means. 

Film-Storage Method The film-storage method of large-screen tele- 
vision projection is the only storage system available even on an ex- 

Fig. 2 Rapid processing unit. 

perimental basis in this country. The system described here was de- 
veloped by Paramount Pictures and has been used on several occasions 
in the Paramount Theater in New York City. While developed by 
Paramount, the fundamentals are similar in many respects to equip- 
ments designed and built by others and may give the motion picture 
industry an insight into the problem involved in setting up such a 

The film-storage system consists of four basic elements : 1 . television 
receiving equipment; 2. recording camera; 3. rapid film-processing 



equipment; 4. a conventional 35-mm motion picture projector. 
Illustrations of such equipment are shown in Figs. 2 and 3. In prac- 
tice, Paramount has used mobile cameras together with microwave 
radio-relay equipment to bring the program material to the theater. 
The mobile cameras with associated control equipment and micro- 
wave-relay unit are of the conventional type used by television broad- 
casters for remote pickup and cost approximately $55,000. 

Fig. 3 Complete equipment in operation. 

Receiver All receiving equipment* is housed in one unit. This in- 
cludes all video and audio equipment together with high- and low- 
voltage supplies. Two screens are provided. One employs a 15-inch 
cathode-ray tube for monitoring; the other is a 10-inch cathode- 
ray tube having an aluminum-backed, flat-face screen. This 10-inch 
cathode-ray tube is of the blue, short-persistence type and provides 
the received image which is photographed. This screen has the po- 
larity reversed and the received image is a negative. Audio portions 
of the program are monitored by a loudspeaker included in this unit. 
* NOTE: The total cost of receiver, camera, and processing unit is approximately 
$35,000 plus installation. 


Camera A special recording camera* is employed having no mechan- 
ical shutter but having its pulldown mechanism synchronized at the 
standard film rate of 24 frames per second with an electronic shutter 
incorporated in the circuits of the 10-inch cathode-ray tube. Twenty 
frames following exposure of the picture, the film passes through the 
sound modulator. A film magazine mounted directly above the 
recording camera holds sufficient unexposed film for two hours' con- 
tinuous recording. 

Processing Exposed film from the recording camera passes through a 
chute directly to a high-speed processing unit.* A maximum of 66 
seconds is required to develop, fix, wash, and dry the exposed film. 
Facilities are provided either to wind the processed film on reels or 
feed it directly to the projectors. The processing unit requires a hot- 
and cold-water supply of approximately 20 gallons per minute. The 
hot-water supply must have a minimum temperature of 140 degrees 
Fahrenheit . Cold-water supply at conventional tap temperature is ade- 
quate. Automatic mixing is provided within the unit to attain a re- 
sultant temperature of approximately 125 degrees Fahrenheit. A slop- 
sink should be provided for disposal of spent photographic chemicals. 
Power The total power required to operate the three units (receiver, 
camera, and processing) is 100 amperes, 3-phase, 208 volts, alternating 

Space Requirements The space required to house the receiving, re- 
cording, and processing units is 200 square feet. To facilitate opera- 
tion and maintenance, a room 10 X 20 feet is recommended with the 
equipment set up in a straight line allowing at least a 2-foot aisle on 
all sides. 

Electronic-Storage Methods Equipment in this category is not cur- 
rently available for use in American theaters and it does not appear 
that such equipment will be available in the immediate future. 

Two basic systems are described, however, in this report. The first 
uses the dark trace or Skiatron types of screens which are known in 
the American market as P-10 phosphors. Manufacturers in this coun- 
try do not plan in the near future to market a tube which has the 
proper characteristics for television, and some of them express the 
opinion that this screen is not feasible for such use. 

This fact is, of course, well known to the industry from the results of 
published research by many independent investigators as well as the 

* NOTE: The total cost of these three units (receiver, camera, and processing 
unit) is approximately $35,000 plus installation. 


engineers from some of the companies contacted in this survey. Gen- 
erally speaking, the Skiatron tube, at the present development in the 
art, produces an image which does not permit sufficient contrast and 
low persistence to compete successfully with phosphorescent screens or 
with photographic emulsions. It is also difficult to produce a screen 
which produces true blacks and whites. Similarly, its decay time is a 
complex phenomenon, and although it can be controlled to some ex- 
tent in manufacture, satisfactory performance in this regard has not 
been obtained to date. It is entirely possible, however, that future 
developments may reverse present thinking in this regard. Any such 
trend will be noted and presented in a future report of this committee. 

The second storage system was developed in Switzerland and is 
known as the AFIF Method of Large-Screen Television Projection. 
This system was developed by Dr. F. Fischer at the Swiss Federal 
Institute of Technology. Since it was known that this system was not 
currently available for sale, no contact was established with this 
Institute. Because the operation of this system is not well known in 
this country, a brief description is given in the Appendix. 

Although not commercially available, a laboratory model of a 
theater projector using this system was demonstrated in Zurich, 
Switzerland, during the week of September 5, 1948. Reports from 
those viewing the demonstration were that screen brightness was 
equivalent to present motion picture practice and picture definitions 
were adequate for theater use. The demonstration was conducted, 
however, using 729 lines rather than 525 now currently standard for 
broadcast purposes in this country. 


Coaxial Cables and Radio Relays The Bell System coaxial cable and 
radio-relay facilities installed and in operation are as follows : 




New York-Washington 4 Philadelphia 

New York-Newark 1 

New York-Boston 2 

Washington-Richmond 1 


Chicago-Cleveland 2 Toledo 

Chicago-St. Louis 2 Danville (pickup only) 

Chicago-Milwaukee 1 

Cleveland-Buffalo 1 

Toledo-Detroit 2 


It is planned to have two channels in operation between Philadel- 
phia and Cleveland by January 1, 1949, with an intermediate ter- 
minal at Pittsburgh, which will join the eastern and midwest net- 

Tentative additional facilities planned for completion in 1949 are as 
follows : 




New York- Washington 1 As required 

Los Angeles-San Francisco 2 

Milwaukee-Madison 1 

Also planned for 1949 are additional connecting facilities to approxi- 
mately six cities on existing routes. 

In 1950 the following routes are planned : 


New York-Boston 3 As required 

New York-Chicago 3 As required 

New York-New Haven 1 


Toledo-Cincinnati Dayton 


Dayton-Louisville 2 Indianapolis 

Philadelphia-Wilmington 1 


Boston-Providence 1 

The additional facilities planned for 1949 and 1950 between New 
York and Washington and between New York and Boston will be of 
the same type now installed between these points. In the case of the 
New York-Chicago route, however, the three additional channels will 
be provided by radio relay. 

At the present time, there are no definite plans for a transconti- 
nental network. Development of such a network or other routes will 
depend upon the needs of the television industry. 

Charts showing the existing and planned facilities of the Bell system 

are shown in Figs. 4 and 5. 


Rates The following are the proposed rates for intercity video chan- 
nel services which are now under review by the Federal Communica- 
tions Commission. 



A. Monthly Service Where Allocation of Usage Is Not Required 

Service Seven Days per Week 

Per airline mile, per months 
Eight consecutive hours or fraction 

thereof per day $ 35.00 per month 

Each additional consecutive hour or 

fraction thereof per day $ 2.00 per month 

Additional Hours, per Occasion of Use 

When the additional hours precede or 
succeed and are consecutive with the 
daily service period : 
Per airline mile, per hour or fraction 

thereof $ 0.25 per hour 

When the additional hours are not con- 
secutive with the daily service period : 
Per airline mile, per hour or fraction 

thereof $ 0.50 per hour 


Service Seven Days per Week 
Each station connection, per month : 
Eight consecutive hours or fraction 

thereof per day $500.00 per month 

Each additional consecutive hour or 

fraction thereof per day $ 35.00 per month 

Additional Hours, per occasion of Use 

When the additional hours precede or 
succeed and are consecutive with the 
daily service period : 
Each connection, per hour or fraction 

thereof $ 5.00 per hour 

When the additional hours are not con- 
secutive with the daily service period : 
Each connection, per hour or fraction 

thereof $ 10.00 per hour 






Monthly Service Where Allocation of Usage Is Required 

When allocation of usage of available facilities is necessary to 
meet the requirements of two or more customers for monthly 
service, such service will be furnished over the facilities involved 
only at the following rates. Subject to thirty days' notice to cus- 
tomers receiving monthly service under A above, such service 
will be terminated when allocation of usage is required. Subject 
to thirty days' notice to customers receiving monthly service 
under the following rates, such service will be terminated when 
allocation of usage is no longer required. Monthly service usage 
of available facilities will be equitably allocated by the Telephone 
Company to meet in so far as practicable the reasonable require- 
ments of all monthly service customers. 


Service Seven Days per Week 

Per airline mile, per month: 

First four hours or fraction thereof per 
day (composed of consecutive or 
nonconsecutive periods in multiples 
of 15 minutes) 

Each additional hour or fraction thereof 
per day (composed of consecutive or 
nonconsecutive periods in multiples 
of 15 minutes) 

$ 25.00 per month 

$ 4.00 per month 

Additional Hours, per Occasion of Use 

When the additional hours are consecutive 

with the daily service period : 
Per airline mile, per hour or fraction 


When the additional hours are not con- 
secutive with the daily service period : 
Per airline mile, per hoar or fraction 

$ 0.25 per hour 

$ 0.50 per hour 







B. Monthly Service Where Allocation of Usage is Required 


Service Seven Days per Week 

Each station connection, per month : 
First four hours or fraction thereof per 
day (composed of consecutive or non- 
consecutive periods in multiples of 
15 minutes) 

Each additional hour or fraction thereof 
(composed of consecutive or noncon- 
secutive periods in multiples of 15 

$350.00 per month 

$ 60.00 per month 

Additional Hours, per Occasion of Use 

When the additional hours are consecutive 

with the daily service period : 
Each connection, per hour or fraction 


When the additional hours are not con- 
secutive with the daily service period : 
Each connection, per hour or fraction 

$ 5.00 per hour 

$ 10.00 per hour 

Maximum Allocation Charges for a Route or Interconnected Routes 

The charge for each customer's allocated service will be the total 
charges computed at these allocated usage rates less the proportion 
that the total allocation charges for all such customers exceeds, if any, 
the charges which would obtain if one customer had used the entire 
service under the regular monthly service rates, A above, includ- 
ing the charges applicable to switches (D below). 

C. Occasional Service Minimum One Hour 

The maximum charge for occasional service will be that for 
monthly service under A above. 


C. Occasional Service Minimum One Hour 


- - 


Per airline mile: 

First hour or fraction thereof $ 1.00 per hour 

Each additional 15 minutes or fraction 

thereof consecutive with the initial 

period $ 0.25 per hour 



Each station connection, per month (plus 
$10.00 per hour of use or fraction 
thereof) $200.00 per month 

D. Switches 

Each switch of a section of a network (ex- 
cept where allocation of service is re- 
quired) $ 1.00 


From the foregoing report, it is clearly evident that theater-tele- 
vision equipment has been developed which is capable of providing 
pictures of continuing entertainment value. While not equal in 
quality to present 35-mm film, evidence has been presented which in- 
dicates such quality will be approached in the future. Methods of 
distribution of program material by coaxial cables or radio channels 
also have reached a stage of development where satsifactory tele- 
vision pictures can be transmitted over necessary distances. 

Further development of equipment as well as provision by the 
Federal Communications Commission of suitable radio channels is 
now mainly dependent upon the interest shown by the motion picture 
industry. Active participation by theater owners and related organi- 
zations is essential if the opportunity to use this new medium is not to 
be lost. 

The FCC, however, does not grant channel allocations on a vague 
request that they may be needed at some future date. Concrete 
evidence must be presented that the group requesting such allocations 
is prepared financially and technically to provide a service in the 




public interest. Only by such action can it be hoped that the request 
will receive favorable consideration. 

The radio-frequency spectrum is very rapidly becoming over- 
crowded. If the motion picture industry ever hopes to use television 
in the theater, action must be taken now. A year from now may be 
too late. Producers, distributors, and exhibitors alike must unite and 
approach the FCC with a well-formulated plan that they seriously 
intend immediate experimental operation. 


AFIF Television Projection System The television projection system 
described below was developed by the Gesellschaft der Forderung der 
Forschung auf dem Gebeite der Technischen Physik an der Eidgenos- 






Fig. 6 

sischen Technischen Hochschule of Zurich, Switzerland. For brevity, 
the system will be referred to simply as the "Swiss System." 

The method by which this system forms a large-scale television 
picture can be explained most easily by first considering the optical 
system shown in Fig. 6. 

Light from the source at the left (an ordinary projection arc lamp) 
passes through the condensing-lens system LI , the baffle plate Bl , a 
second lens L2, a transparent plate G (which will be seen to act as a 
light valve), a second baffle plate B2, a projection-lens system L3, and 
finally reaches the screen at the right. 

For simplicity, the optical system can be considered to consist of 
the following 3 component systems : 




1. The light source and LI provide uniform illumination of the 
glass plate G. 

2. The lens L2 forms an image of Bl in the plane of B2 (assuming 
for the moment that G is a glass plate with parallel surfaces) with a 
magnification of unity. The baffle plates Bl and B2 are formed as 
shown in Fig. 7. 

They consist of a series of narrow opaque bars separated by gaps 
equal to the width of the bars. The two baffles are so aligned that the 
images of the gaps in Bl fall on the bars of B2. The result is that no 
light passes through the system as so far described. 

3. The projection lens L8 forms an enlarged image of the plate G 
on the projection screen at the right of Fig. 6. Of course, since no 

light, was transmitted past B8, the 
screen is uniformly dark except for 
a certain unavoidable amount of 
scattered light. 

It has been assumed that the 
plate G consists of a sheet of glass 
with parallel surfaces. Actually 
this is not the case. The construc- 
tion and manipulation of this plate 
are the vital elements of the entire 
system. To understand the func- 
tion of the plate consider the effect 
of a small irregularity on its sur- 
face (the nature of this surface will 



Fig. 7 

be described later) . Let the irregularity be restricted to an area A , 
which has the dimensions of a television-picture element. Further 
let the irregularity have the form of a slight bulge in the surface, the 
height of the bulge being denoted by H . 

A light ray which passes through A will be deflected from its path 
due to the curvature of the surface resulting from the bulge. Some of 
the rays will be deflected enough to clear the bars in the barrier B2 and 
enter the projection system. By increasing the curvature of the sur- 
face (increasing the value of H), the fraction of the total light passing 
through A which is deflected into the projection system is increased. 

It should be noted here that rays which are deflected in a direction 
parallel to the bars of the barrier are still stopped by the barrier, while 
rays deflected at right angles to the bars have an excellent chance of 
getting through. For this reason, a cylindrical distortion of A with 


the axis of the cylinder parallel to the bars is more efficient than a 
spherical distortion since the sphere produces deflection in all direc- 
tions and the cylinder causes deflection only in the preferred 

Since the projection lens forms an image of the plate on the screen, 
all light passing through A and through the gaps in B2 is focused into 
an image area A' on the screen. Further, the brightness of A' is 
roughly proportional to H, the amplitude of the distortion of A. 

At this point it is clear that a television picture can be formed on the 
screen if the surface of G can be divided into a sufficient number of 
small elements, each of which is distorted in cylindrical fashion with an 
amplitude of distortion proportional to the brightness of the corre- 
sponding element of the original scene. This effect is achieved as 
follows : 

The plate consists of a solid sheet of transparent conducting ma- 
terial coated with a liquid layer. An electrostatic charge is placed on 
the surface of the liquid by mounting the entire plate in a vacuum 
tube and scanning it in normal television fashion with a beam of 
electrons from an electron gun which is provided with suitable deflec- 
tion and intensity-control devices. Electrostatic forces between the 
conducting-base plate and the charges thus placed on the surface of 
the liquid cause a deformation of the surface. By somewhat com- 
plicated manipulation of the electron beam this deformation can be 
made to take on the required form outlined above. 

Now consider the cycle of events at one element of the surface. The 
scanning beam passes over the element and establishes the proper 
charge distribution. The liquid surface flows into a corresponding 
distortion pattern and light passes through to the screen. As long as 
this distortion remains, light reaches the screen so that the tube is 
effectively a storage device. However, the time duration of the 
storage must be limited to something less than a scan period so that a 
new value of distortion can be established when the scanning beam 
returns. To this end the liquid is made slightly conducting so that the 
charge variation slowly smooths itself out and the distortion dis- 
appears. Then the beam returns and the cycle repeats. In practice, 
the conductivity is such that the distortion amplitude drops to about 
10 per cent of its peak value in the period of one scan. This gives an 
effective storage period of perhaps one half the scan period. 

The liquid used to coat the plate must also meet other require- 
ments. Since the plate is in a vacuum, a very low vapor pressure is 


necessary. Also the electrostatic forces on the liquid tend to set up a 
flow of material from the center of the scanned area to the edges. Thus 
a rather high viscosity is needed to resist these forces. Such a liquid 
is obtained by mixing apiezon oil, Canada balsam, and a conducting 

In addition, the liquid layer is continually smoothed and cooled by 
using a disk for the base plate and rotating the disk slowly and con- 
tinuously. Thus the scanned area is removed from the scanning posi- 
tion, passed through smoothing and cooling devices, and returned to 
the scanning position. This motion is very slow so that no blurring of 
the image is produced. 

This system is to some extent similar to the Scophony system. For 
one thing, both use "schlieren" optics, that is, optical systems which 
transmit light only when some disturbance is set up in a transmitting 
medium. In both cases the transmitting medium is liquid, but the 
nature of the disturbance is quite different. The Scophony system 
uses a compression wave which travels through a liquid cell. Thus 
the storage period is limited to the travel time of the compression 
wave through the cell, and can never be longer than the period of one 
television line. Also, rotating mirrors are an essential part of the 
Scophony system. 

The Swiss method creates a surface distortion such that each ele- 
ment of the surface is responsible for just one element of the image. 
Theoretically, this would permit storage during the period of a tele- 
vision frame, and in practice storage periods of about one half this 
value are realized. 

The result is a much higher optical efficiency than can be realized 
with the Scophony system. In fact, the screen illumination that can 
be produced is stated to be about 30 per cent of that obtained with a 
motion picture projector using the same light source and projection 





This work has been prepared as an interim report of the Theater 
Television Committee of the Society of Motion Picture Engi- 
neers. It is a statement of the present state of the art written in 
nontechnical language. Information on government regulations, 
types of equipment available, and approximate costs has been in- 
cluded. It is directed to those of the motion picture industry who 
\vish to take advantage of this rapidly expanding entertainment me- 
dium. The membership of the committee directly responsible for 
this work is as follows : 

D. E. HYNDMAN, Chairman 
Eastman Kodak Company 


RCA Victor Division 

F. E. CAHILL, Jr. 
Warner Brothers Pictures 

Ansco Division 


United Photo Supply Corporation 


General Precision Laboratories 


Paramount Pictures 

Consulting Engineer 


Allen B. Du Mont Laboratories 

RKO Theaters 


Loew's, Inc. 

Bell Telephone Laboratories 

Los Alamos Scientific Laboratory 

Warner Brothers Pictures 

Paramount Pictures 

Eastman Kodak Company 

E. I. du Pont de Nemours and Co. 

National Theater Supply Company 

Twentieth Century-Fox Film Cor- 

RCA Victor Division 



(1) William H. Forster, "6000 me television relay system," Electronics, vol. 
22, pp. 80-85; January, 1949. 

(2) Fred G. Albin, "Sensitometric aspect of television monitor tube photog- 
raphy," /. Soc. Mot. Pict. Eng., vol. 51, pp. 595-613; December, 1948. 

(3) Richard Hodgson, "Paramount Pictures' system of theater television," 
Presented SMPE 64th convention, Monday, October 25, 1948, /. Soc. Mot. PicL 
Eng., to be published. 

(4) Roy Wilcox and Hubert J. Schafly, "Theater installation of instantaneous 
large-screen television," Presented SMPE 64th convention, Monday, October 25, 
1948, J. Soc. Mot. Pict. Eng., to be published. 

(5) Ernest H. Schreiber, "Video distribution facilities for television trans- 
mission," /. Soc. Mot. Pict. Eng., vol. 51, pp. 574-589; December, 1948. 

(6) Otto H. Schade, "Electro-optical characteristics of television systems," 
RCA Rev., vol. 9, pp. 5-37, March, 1948; pp. 246-286, June, 1948; pp. 491-530, 
September, 1948; pp. 653-686, December, 1948. 

(7) J. L. Boon, W. Feldman, and J. Stoiber, "Television recording camera," 
/. Soc., Mot. Pict. Eng., vol. 51, pp. 117-126; August, 1948. 

(8) Thomas T. Goldsmith, Jr., and Harry C. Milholland, "Television trans- 
scription by motion picture film," /. Soc. Mot. Pict. Eng., vol. 51, pp. 107-116; 
August, 1948. 

(9) A. G. D. West, "Development of theater television in England," J. Soc. 
Mot. Pict. Eng., vol. 51, pp. 127-168; August, 1948. 

(10) H* L. Logan, "Brightness and illumination requirements," /. Soc. Mot. 
Pict. Eng., vol. 51, pp. 1-12; July, 1948. 

(11) Ralph V. Little, Jr., "Developments in large-screen television," J. Soc. 
Mot. Pict. Eng., vol. 51, pp. 37-51; July, 1948. 

(12) I. G. Maloff, "Optical problems in large-screen television," J. Soc. Mot. 
Pict. Eng., vol. 51, pp. 30-36; July, 1948. 

(13) Robert M. Fraser, "Motion picture photography of television images," 
RCA Rev., vol. 9, pp. 202-217; June, 1948. 

(14) Boyce Nemec, "Review of SMPE work on screen brightness," J. Soc. 
Mot. Pict. Eng., vol. 50, pp. 254-259; March, 1948. 

(15) "Report of Screen Brightness Committee," /. Soc. Mot. Pict. Eng., vol. 
50, pp. 260-273; March, 1948. 

(16) Alfred N. Goldsmith, "Theater television A general analysis," J. Soc. 
Mot. Pict. Eng., vol. 50, pp. 95-121; February, 1948^ 

(17) F. M. Deerhake, "2000 me television program chain," Electronics, vol. 21, 
pp. 94-97; February, 1948. 

(18) Harry R. Lubcke, "Effect of time element in television program opera- 
tions," J. Soc. Mot. Pict. Eng., vol. 48, pp. 543-547; June, 1947. 

(19) L. G. Abraham and H. I. Romnes, "Television network facilities," 
Elec. Eng., vol. 66, p. 47; May, 1947. 

(20) "Statement of SMPE on revised frequency allocations," /. Soc. Mot. 
Pict. Eng., vol. 48, pp. 183-202; March, 1947. 



(21) W. Boothroyd, "UHF television relay system," Electronics, vol. 20, p. 
86; March, 1947. 

(22) Lester B. Isaac, "Television and the motion picture theater," /. Soc. 
Mot. Pict. Eng., vol. 47, pp. 482-486; March, 1947. 

(23) W. J. Poch and J. P. Taylor, "Microwave equipment for television relay 
services," Broadcast News, vol. 44, p. 20; October, 1946. 

(24) Allen B. Du Mont, "The relation of television to motion pictures," J. Soc. 
Mot. Pict. Eng., vol. 47, pp. 238-247; September, 1946. 

(25) Paul J. Larsen, "Report of Committee on Television Projection Practice," 
/. Soc. Mot. Pict. Eng., vol. 47, pp. 118-119; September, 1946. 

(26) C. F. White and M. R. Boyer, "A new film for photographing the tele- 
vision monitor tube," J. Soc. Mot. Pict. Eng., vol. 47, pp. 152-164; August, 1946. 

(27) Ralph B. Austrian, "Film The backbone of television programming," 
/. Soc. Mot. Pict. Eng., vol. 45, pp. 401-413; December, 1945. 

(28) A. H. Rosenthal, "Problems of theater television projection equipment," 
/. Soc. Mot. Pict. Eng., vol. 45, pp. 218-240; September, 1945. 

(29) "Frequency allocations for theater television," J. Soc. Mot. Pict. Eng., 
vol. 45, pp. 16-19; July, 1945. 

(30) D. W. Epstein and I. G. Maloff, "Projection television," J. Soc. Mot. 
Pict. Eng., vol. 44, pp. 443-455; June, 1945. 

(31) Harold S. Osborne, "Coaxial cables and television transmission," J. Soc. 
Mot. Pict. Eng., vol. 44, pp. 403-418; June, 1945. 

(32) Ralph B. Austrian, "Some economic aspects of theater television," J. Soc. 
Mot. Pict. Eng., vol. 44, pp. 377-385; May, 1945. 

(33) J. Huber, "Theoretical inquiries on distortions undergone by electrical 
television signals in vacuum-tube connections," Schweizer Archiv., vol. 11, p. 77, 
March, 1945; p. 115, April, 1945. 

(34) "Report of the British Television Committee 1943," London, His Majes- 
ty's Stationery Office, 1945. 

(35) "Statement of the SMPE in opposition to the brief of the CBS as it re- 
lates to theater television," /. Soc. Mot. Pict. Eng., vol. 14, pp. 263-274; April, 

(36) "Statement of the SMPE on allocation of frequencies in the radio fre- 
quency spectrum from 10 kilocycles to 30 million kilocycles for theater television 
service," /. Soc. Mot. Pict. Eng., vol. 44, pp. 105-117; February, 1945. 

(37) Paul J. Larsen, "Statement presented before the Federal Communica- 
tions Commission relating to television broadcasting," /. Soc. Mot. Pict. Eng., 
vol. 44, pp. 123-127; February, 1945. 

(38) Alfred N. Goldsmith, "Future of theater television," Television, Febru- 
ary, 1945. 

(39) Ralph R. Beal, "Electronic research opens new frontiers," Proc. I.R.E., 
vol. 33, pp. 5-9; January, 1945. 

(40) T. M. C. Lance, "Some aspects of large screen television," /. Telev. Soc., 
vol. 4, p. 82; December, 1944. 

(41) L. de Forest, "Early beginnings of large screen television," Program, 
First Annual Conference, Television Broadcasters Association, p. 37; December 
U-12, 1944. 


(42) Paul J. Larsen, "Theatrical television," Program, First Annual Confer- 
ence, Television Broadcasters Association, p. 39; December 11-12, 1944. 

(43) Paul J. Larsen, "Panel meeting on theater television," Proceedings of 
First Annual Conference, Television Broadcasters Association, p. 108; December 
11-12, 1944. 

(44) E. K. Carver, R. H. Talbot, and H. A. Loomis, "Film distortions and their 
effect upon projection quality," J. Soc. Mot. Pict. Eng., vol. 41, pp. 88-93; July, 

(45) E. E. Masterson and E. W. Kellogg, "A study of flicker in 16-millimeter 
picture projection," J. Soc. Mot. Pict. Eng., vol. 39, pp. 232-244; October, 1942. 

(46) P. C. Goldmark, J. N. Dyer, E. R. Piore, and J. M. Hollywood, "Color 
television," J. Soc. Mot. Pict. Eng., vol. 38, pp. 311-352; April, 1942. 

(47) M. R. Null, W. W. Lozier, and D. B. Joy, "Color of light on the projec- 
tion screen," /. Soc. Mot. Pict. Eng., vol. 38, pp. 219-228; March, 1942. 

(48) F. Fischer and H. Thiemann, "Theoretical considerations on a new proc- 
ess for television large-screen projection," Schweizer Archiv., vol. 7, p. 1, January, 
1941; p. 33, February, 1941; p. 305, November, 1941; p. 337, December, 1941. 

(49) "Television report Order, rules, and regulations," J. Soc. Mot. Pict. 
Eng., vol. 37, pp. 87-97; July, 1941. 

(50) I. G. Maloff and W. A. Tolson, "A resume" of the technical aspects of 
RCA theater television," RCA Rev., vol. 6, p. 5; July, 1941. 

(51) "Baird high definition color television," /. Telev. Soc., vol. 3, p. 171; 

(52) E. A. Williford, "Twenty-four years of service in the cause of better pro- 
jection," /. Soc. Mot. Pict. Eng., vol. 36, pp. 294-301; March, 1941. 

(53) C. Frederick Wolcott, "Problems in television image resolution," /. Soc. 
Mot. Pict. Eng., vol. 36, pp. 65-81; January, 1941. 

(54) "Report of Television Committee," J. Soc. Mot. Pict. Eng., vol. 35, pp. 
569-583; December, 1940. 

(55) F. H. Richardson, "Advancement in projection practice," J. Soc. Mot. 
Pict. Eng., vol. 35, pp. 466-483; November, 1940. 

(56) M. W. Baldwin, "The subjective sharpness of simulated television 
images," Proc. I.R.E., vol. 28, pp. 458-168; October, 1940. 

(57) P. C. Goldmark and J. N. Dyer, "Quality in television pictures," /. Soc. 
Mot. Pict. Eng., vol. 35, pp. 234-253; September, 1940. 

(58) "A new projection system for large-screen television receivers," Elec- 
tronics and Telev. and Short-Wave World, vol. 13, p. 372; August, 1940. 

(59) W. C. Kalb, "Progress in projection lighting," J. Soc. Mot. Pict. Eng., 
vol. 35, pp. 17-31; July, 1940. 

(60) A. H. Rosenthal, "A system of large-screen television reception based on 
certain electron phenomena in crystals," Proc. I.R.E., vol. 28, pp. 203-213; May, 

461) F. Fischer, "On the paths to television large-screen projection," Schweizer 
Archiv., vol. 6, p. 89; April, 1940. 

(62) Alfred N. Goldsmith, "Future development in the field of the projection- 
ist," J. Soc. Mot. Pict. Eng., vol. 34, pp. 131-142; February, 1940. 

(63) T. Mulert, "Viewpoints in the construction of television projection tube 
receivers," Fernseh A. G. Hausmitteilungen, vol. 1, p. 217; December, 1939. 


(64) G. Schubert, W. Dillenburger, and H. Zschau, "The intermediate film 
process," Fernseh A. G. Hausmitteilungen, vol. 1, Part I, p. 65; April, 1939. 
Part II, p. 162; August, 1939. Part III, p. 201; December, 1939. 

(65) "The latest Scophony big-screen projector," Electronics and Telev. ami 
Short-Wave World, vol. 12, p. 654; November, 1939. 

(66) W. C. Harcus, "Screen color," /. Soc. Mot. Pict. Eng., vol. 33, pp. 444- 
148; October, 1939. 

(67) Lorin D. Grignon, "Flicker in motion pictures," J. Soc. Mot. Pict. Eng., 
fol. 33, pp. 235-247; September, 1939. 

(68) D. M. Robinson, "The supersonic light control," Proc. I.R.E., vol. 27, 
D. 483; August, 1939. 

(69) J. Sieger, "The design and development of television receivers using the 
Scophony television optical scanning system," Proc. I.R.E., vol. 27, pp. 487-492; 
Yugust, 1939. 

(70) G. Wikkenhauser, "Synchronization of Scophony television receivers," 
roc. I.R.E., vol. 27, pp. 492^96; August, 1939. 

(71) H. W. Lee, "Some factors involved in the optical design of a modern 
/elevision receiver using moving scanners," Proc. I.R.E., vol. 27, pp. 496-501; 
August, 1939. 

(72) R. Moeller and G. Schubert, "Further developments in receiver and 
jicture pickup equipment in the year 1939," Fernseh A. G. Hausmitteilungen, 
ro\. 1, p. 153; August, 1939. 

(73) C. H. Bell, "Facts and problems for the cinema," Telev. and Short-Wave 
World, vol. 12, p. 467; August, 1939. 

(74) R. Moeller and G. Schubert, "Ten years of television technique," Fernseh 
4. G. Hausmitteilungen, vol. 1, p. Ill; July, 1939. 

(75) E. Schwartz, "The development of television Braun tubes by Fernseh 
V. G.," Fernseh A. G. Hausmitteilungen, vol. 1, p. 123; July, 1939. 

(76) Allen B. Du Mont, "Design problems in television systems and receiv- 
es," J. Soc. Mot. Pict. Eng., vol. 33, pp. 66-74; July, 1939. 

(77) "Report of Television Committee," /. Soc. Mot. Pict. Eng., vol. 33, pp. 
'5-79; July, 1939. 

(78) "The EMI cinema projector," Telev. and Short-Wave World, vol. 12, p. 
J89; July, 1939. 

(79) "Report of Projection Practice Committee," /. Soc. Mot. Pict. Eng., 
rol. 33, pp. 101-105; July, 1939. 

(80) O. Reeb, "A consideration of the screen-brightness problem," J. Soc. 
Mot. Pict. Eng., vol. 32, pp. 485-494; May, 1939. 

(81) R. Moeller, "The lens array screen," Fernseh A. G. Hausmitteilungen, 
rol l,p.73; April, 1939. 

(82) "Baird cinema equipment," Telev. and Short-Wave World, vol. 12, p. 
L99; April, 1939. 

(83) "Kolorama television" Telev. and Short-Wave World, vol. 12, p. 137; 
Vlarch, 1939. 

(84) "Making history (circumstances of the televising of the Boon-Danaher 
>rize fight)," Telev. and Short-Wave World, vol. 12, p. 131; March, 1939. 

(85) P. W. S. Valentine, "The Philips large screen television receiver," J. 
Telev. Soc., vol. 3, p. 7; January, 1939. 


(86) "Report of the Projection Practice Committee," J. Soc. Mot. Pict 
Eng., vol. 31, pp. 480-510; November, 1938. 

(87) E. Schwartz, H. Struebig, and H. W. Paehr, "Beam generation in tele- 
vision cathode-ray tubes for projection purposes," Fernseh A. G. Hausmitteil-l 
ungen, vol. 1, p. 5; August, 1938. 

(88) W. Dillenburger, "Radio large projection receiver with 1.5 square meter 1 
picture surface," Fernseh A. G. Hausmitteilungen, vol. 1, p. 29; August, 1938. 

(89) "Report of the Projection Practice Committee," /. Soc. Mot. Pict. Eng., 
vol. 30, pp. 636-650; June, 1938. 

(90) F. H. Richardson, "A discussion of screen-image dimensions," J. Soc. 
Mot. Pict. Eng., vol. 30, pp. 334-338; March, 1938. 

(91) "Report of the Projection Practice Committee," /. Soc. Mot. Pict. 
Eng., vol. 29, pp. 614-621; December, 1937. 

(92) A. Goetz and W. O. Gould, "The objective quantitative determination 
of the graininess of photographic emulsions," J. Soc. Mot. Pict. Eng., vol. 29, pp. ^ 
510-538; November, 1937. 

(93) V. K. Zworykin and W. H. Painter, "Development of the projection ; 
kinescope," Proc. I.R.E., vol. 25, pp. 937-954; August, 1937. 

(94) R. R. Law, "High-current electron gun for projection kinescope," f 
Proc. I.R.E., vol. 25, p. 954-977; August, 1937. 

(95) "Report of the Projection Screen Brightness Committee," J. Soc. Mot. 
Pict. Eng., vol. 27, pp. 127-139; August, 1936. 

(96) "Report of the Projection Screen Brightness Committee," /. Soc. Mot 
Pict. Eng., vol. 26, pp. 489-504; May, 1936. 

(97) R. P. Teele, "Photometry and brightness measurements," J. Soc. Mot. \ 
Pict. Eng., vol. 26, pp. 554-569; May, 1936. 

(98) S. K. Wolf, "An analysis of theater and screen illumination data," J. 
Soc. Mot. Pict. Eng., vol. 26, pp. 532-542; May, 1936. 

(99) M. Luckiesh and F. K. Moss, "The motion picture screen as a lighting 
problem," J. Soc. Mot. Pict. Eng., vol. 26, pp. 578-591; May, 1936. 

(100) W. F. Little and A. T. Williams, "Resumd of methods of determining 
screen brightness and reflectance," J. Soc. Mot. Pict. Eng., vol. 26, pp. 570-577; 
May, 1936. 

(101) C. M. Tuttle, "Density measurements of release prints," J. Soc. Mot. 
Pict. Eng., vol. 26, pp. 548-553; May, 1936. 

(102) A. A. Cook, "A review of projector and screen characteristics, and their 
effects upon screen brightness," /. Soc. Mot. Pict. Eng., vol. 26, pp. 522-531; 
May, 1936. 

(103) B. O'Brien and C. M. Tuttle, "An experimental investigation of projec- 
tion screen brightness," J. Soc. Mot. Pict. Eng., vol. 26, pp. 505-521; May, 1936. 

(104) E. M. Lowry, "Screen brightness and the visual functions," J. Soc. Mot. 
Pict. Eng., vol. 26, pp. 490-504; May, 1936. 

Research Council 
Small Camera Crane 



Summary The Research Council small camera crane was designed from 
requirements and specifications set down by the Council's Photographic 
Committee. While similar in principle to other cranes, it embodies an en- 
tirely new design and built-in safety features. The crane dolly is electrically 
driven. The boom arm is manually operated and can be panned through 
360 degrees. The crane is large enough to have a lens height of from 2 to 
10 feet from floor level and small enough so that, fully equipped, it will pass 
through a doorway 6 feet high and 36 inches wide. 

THIS CRANE, known as the Research Council small camera crane, 
is similar in principle to other cranes used in the motion picture 
industry, but it embodies an entirely new design and built-in safety 
features. It can be built as a standard production piece of equipment 
at a reasonable price. 

The specifications of the Camera Crane Committee, composed of 
members from each studio, called for a self-propelled, remote-con- 
trolled crane, large enough to have a lens height of from 2 to 10 feet 
from the floor level, but small enough so that fully equipped it would 
pass through a doorway 36 inches Avide and 6 feet high. The crane 
was designed to accommodate the Technicolor equipment; the 
weight to be of minimum weight possible in order to permit its use on 
the present stage floors, preferably without track; and deflection and 
distortion to be kept to a minimum. 

From the specifications, the present Research Council crane was 
designed and built with the use of aluminum alloy wherever possible, 
in order to keep weight to a minimum and castings were used to keep 
deflections to a minimum. After load-testing it thoroughly, we ob- 
served that deflection under weight was negligible. 

Although the cost of patterns was high, we have shown this method 
to be the best and most economical way of building a production 
boom. The weight of the crane itself is about 1200 pounds. 

The drive unit is actuated by a 2-horsepower, 110-volt, direct- 

* Presented May 18, 1948, at the SMPE Convention in Santa Monica. 





current, series-wound motor, especially designed for this crane by the 
General Electric Company. This motor is supported on rubber 
mounts and is coupled to the differential drive by a more-flex coupling. 
The differential is a 10: 1 cone worm-drive unit, with all driving sur- 
faces lapped in assembly. This design was considered necessary to 
insure smooth and silent operation. The differential housing is a full- 
floating axle type and the full weight of the crane is carried by the 
housing itself, leaving the driving axles torque-loaded only. 

The motor can be remotely controlled if desired. The control itself 
is a complete and separate unit consisting of resistors, reversing 
switch, and solenoid brake. With this control, various degrees of 

Fig. 1 Drive unit. 

acceleration or deceleration can be obtained at will. A Cannon plug 
is used to connect the control to the crane. 

The brake is a friction, air-cooled, disk-type, actuated by a 110-volt, 
direct-current solenoid. 

The brake solenoid is actuated by a carbon pile to provide a sort of 
"feel control" which gives the operator a proportional amount of 
braking power to the pressure applied to the carbon pile. 

The steering unit is of an unusual design which permits the crane 
to be completely revolved within a 6-foot radius. It also allows the 
crane to be placed squarely against a wall with the least maneuvering. 
This device is provided with a lock-preventing arm, allowing a very 
sharp turn with the least amount of effort. 

The following units are mounted on the base casting : (a) the drive 
unit, (b) the steering unit, (c) the center post, (d) the hydraulic pump 
and take-up tank, and (e) four jacks which are interconnected to 






Fig. 2 Steering unit. 

permit the whole crane to be lifted 
for service or to protect the wheels 
during storage periods. 

The center post is a 7-inch tele- 
scoping tube mounted on ball bear- 
ings. It permits the boom to be 
panned 360 degrees around the 
crane base, tilted up 55 degrees, 
and down 45 degrees from the 
horizontal position. The boom 
arm and parallel bar are mounted 
on it. A hydraulic cylinder of 
16-inch extension is mounted in 
the center of the telescoping tube. 
A flow restrictor is located on the 

cylinder base, providing added safety to equipment and personnel, 
as it limits the downstroke to a predetermined speed in case of acci- 
dental breaking of any hydraulic line or mishandling of equipment. 

The panning brake is hand-operated by moving a lever in either 
direction from the centerline and can be adjusted to any degree of 
friction required by the operator, but 
cannot be locked rigidly. This is to pre- 
vent damage to the crane in case of ac- 
cidental shock to the boom. A special 
positioning device operates automati- 
cally to keep a uniform degree of friction ; 
another automatic locking pin prevents 
using the panning brake when the hy- 
draulic pump is in use or the use of the 
hydraulic pump when the panning brake 
is not in the neutral position. 

The tilting brake is located in the 
center post and the boom casting itself is 
relieved to form the brake drums. The 
brake is hand-operated. Handles are 
provided on each side of the boom 
and can be set to any degree of fric- 
tion. Adjusting screws are pro- 
vided to limit the maximum braking 
force, Fig. 3 Center post, 







Fig. 4 Panning brake. 

The camera platform is mounted on the forward or long end of the 
boom. It is kept in a constant horizontal position by the parallel 
bar. The platform is machined to receive the camera table, the 
camera-table-leveling device, and a series of common and twist plugs. 
These are to accommodate cables for the camera motor, the remote- 
focusing device, and lights needed by the camera crew. 

The counterweight inner and outer boxes are mounted on the rear or 
short end of the boom. The outer box is mounted directly on the 
boom and parallel bar, and the inner base is closely mounted in the 
interior of the outer box. The inner box is retractable and can be 
extended out about 16 inches which gives a greater degree of safety 
for heavy loads such as the Technicolor equipment. This provides 
extra room for lead counterweights if needed. 













Fig. 5 Tilt brake, 






Fig. 6 Counterweight boxes. 

The Vernier counterweight is a hand-operated device which can be 
operated from each side of the boom and corrects for an unbalance up 
to 35 pounds. It is located in the shell of the boom. 

The safety -tilt device is entirely enclosed in the shell of the forward 
or long end of the boom. It is composed of a 2-inch hydraulic cylinder 
anchored on one end to the center-post casting and on the other end 
to the boom itself. Any vertical change of position of the boom 
shortens or lengthens the distance between these two points, causing a 
displacement of fluid from one end of the cylinder to the other. This 
flow is controlled in one direction only; i.e., on the upward movement, 
thus allowing the cameraman or his assistant to leave their respective 
seats without endangering the balance of the boom. 

The control of this device is accomplished by the use of a hydraulic 
solenoid valve which, in turn, is energized whenever the cameraman 
or his assistant or both leave their seats. All action returns to normal 





Fig. 7 Vernier counterweight. 




when the seats are reoccupied or an equal amount of counterweight is ! 
removed from the opposite end of the boom. The pressure-relief J 
valve, which is part of this safety-tilt mechanism, goes into action, 
whenever too much weight is removed from the camera table. As the* 
boom is allowed to rise slowly, this prevents a sudden drop of thel 
boom arm or an unsafe unbalance of the crane. 

The camera-turret table is an integral part of the crane and is per- 
manently mounted on the camera platform. It consists of a camera- 
leveling table allowing 7 degrees of correction in all positions. On 
this table the following are mounted : 








Fig. 8 Safety-tilt mechanism. 

(a) An operator's seat, capable of 360-degree rotation, with ad- 
justable height, length, and back rest. 

(b) An assistant's seat with two possible locations permitting the 
focusing of the lens from either side of the camera. 

(c) A compensating counterweight to offset the natural cen- 
trifugal effect during a fast-panning shot. 

(d) A two-speed panning mechanism which will carry the turret 
fully equipped to any position and at any chosen acceleration. 

(e) Two-gear head adapters for use with all existing gear heads. 

(f) Means are provided so that the camera-turret table and the 
camera can be rotated as a unit or separately, permitting an overshot 
and correcting the camera position without the necessity of moving 




the cameraman and his assistant, therefore giving a much smoother 

(g) A positive camera turret lock at easy reach of the cameraman's 
hand. It can be set to any degree of friction desired. 

When the Research Council camera crane was designed, it was 
foreseen that the crane could and would be used off the stage; that is, 
on street sets or on location. As a single piece of equipment could not 
meet this and the other requirements, it was found necessary to design 
a subchassis which is now called a "trailer." When the crane is 
mounted on the trailer, it is elevated 4 x / 2 inches from the floor level. 

Fig. 9 Camera- turret table. 

The trailer is motorized by direct friction drive from the crane unit, 
thus eliminating the necessity of a second driving unit. It is equipped 
with wheel brakes, either hand- or foot-operated, a drive seat for the 
crane operator on each side of the trailer, and a platform for the boom 

No equipment other than the crane and trailer is necessary to com- 
bine the two for use. The assembling can be done by two men in less 
than 5 minutes. A towing attachment is provided with each trailer 
for transportation to location. 

Experiment in Stereophonic Sound* 



Summary This paper reports an extension of the theory and methods 
of stereophonic recording and reproduction, as particularly applicable to 
motion pictures. Microphone technique becomes very different from that 
previously used because of the manner of staging, the use of varied angles of 
view by the camera, and a fixed theater picture size. Typical microphone 
technique is illustrated and re-recording with added sound effects is de- 
scribed. Resulting conclusions and observations establish a good foundation 
for further work in this field. 


HE THEORETICAL BASIS of stereophonic recording and reproduction 
.L is rather generally known. However, for convenience, it will be 
restated as follows: If an infinite number of ideal microphones could 
be placed within a three-dimensional region bounding a source of 
sound energy, each microphone being connected to a distortionless 
transmission system terminated in an ideal reproducer at some other 
location in surroundings identical to those at the source location and 
in space relation to each other as their corresponding microphones, 
then an observer at the distant location would experience the same 
sensation as an observer at the source point. The first compromise to 
the ideal situation considers microphones and reproducers in two- 
dimensional space as an acoustically transparent curtain between the 
source and the observers. The second compromise employs an infi- 
nite number of microphones in a single straight line. It has been found 
that three complete systems give a good subjective approximation 
when three microphones are equally spaced along some straight line 
in relation to the source. Complete descriptions of experiments in 
this direction are given by a Bell Telephone Laboratories mono- 
graph. 1 

The work reported here is an extension of the theory and methods 
for the use of stereophonic sound in motion pictures. The opportu- 
nity to investigate this possibility came about by a desire on the part 
of Twentieth Century-Fox management to evaluate possible technical 

* Presented May 18, 1948, at the SMPE Convention in Santa Monica. 


improvements in motion pictures. Western Electric Company co- 
operated through Electrical Research Products Division by supplying 
film recording and reproducing equipment and other technical 

Methods were devised for recording dialog and music for use in 
motion pictures, without basically changing accepted fundamental 
forms which include the use of long, medium, and close shots and inter- 
cutting techniques. This is not to say that present cutting philoso- 
phy for stereophonic motion pictures is entirely suitable, as there is 
evidence that indicates some new approach is needed. Re-recording, 
with added sound effects, prescore, and playback methods were all 
used. The end result of the experiment to be described was the pro- 
duction of two single-act plays, several full-orchestra numbers (one 
with picture), and a vocal rendition with accompanying orchestra. 

It was concluded that stereophonic methods, with suitable modifica- 
tions, can be applied to motion picture technique and result in a sound 
presentation considerably superior to methods now in use. 


The problem of pickup, particularly for dialog, was first approached 
by setting up a three-channel monitoring system, using the amplifier 
and horn apparatus as installed in the test theater, and providing 
microphones and mixer equipment in an adjacent stage. On this 
stage a small living room set was constructed and pickup tests were 
begun using stock players from the studio roster. 

It was natural first to try the accepted method of having three 
microphones equally spaced and placed in some straight line in rela- 
tion to the actors. This method failed immediately the reasons 
being as follows : 

1. The action was taking place in a restricted space. 

2. Seminondirectional microphones were essentially useless be- 
cause of the proximity to the sources and to acoustic reflections from 
parts of the set which produced false apparent origins. 

3. Actors generally play to other actors and do not face an audi- 
ence as do public speakers. Methods were needed for giving sound 
placement to actors who are speaking at right angles to the camera 
axis and within a few feet of one another. 

4. Since various camera lenses are. used to give emphasis or local- 
ize action, magnifications or distance distortion exists and a similar 
effect was necessary in the sound pickup. 




Using adjustable directivity unidirectional microphones and sepa- 
rate microphone booms for all further work, several basic microphone 
setups resulted. 

In all of the following illustrations, actors will be depicted by a "V" 
within a circle, the apex of the letter indicating the speaking direction; 
microphones by a circle with the protruding arrowhead indicating the 
direction of maximum pickup and the extended tail the direction of 
minimum reception, the inscribed letter stating the connected channel 
as left L, center C, or right R; movement of actors or microphones by 
dashed lines with arrowheads giving direction. 

Fig. 1 

-Typical microphone arrangement for large-set long shot with broad 
actor movement. 

For purposes of definition, consider a source which emits sound 
continuously and which is in constant movement, then the reproduced 
sound must also move continuously in the same way without obvious 
dwells or jumps from position to position and this characteristic will 
be termed smooth sound placement transition. 

The usual equidistant, in-line microphone technique can be used in 
long shots of large sets with wide separation of actors and broad move- 
ments. Even then, to have smooth dialog transition, some micro- 
phone movement may be required, as illustrated in Fig. 1. 

This method is also generally used for recording of effects. One 




experiment using such a pickup consisted in recording airplane take- 
offs, wherein the microphones were placed along the runway. The 
microphones were spaced 150 feet apart! 

When actors are not disposed closer than six to seven feet and are 
speaking directly to each other, the setup of Fig. 2 is used. Should 
either turn or move away and speak lines, then microphones must be 
moved accordingly. For example, should the right-hand actor turn 
180 degrees and speak lines, then the R microphone must be moved 
sufficiently to the right to give good pickup and the C microphone 
readjusted to a somewhat central position, probably favoring the 
right-hand actor. Also, dependent upon set conditions, it is some- 

Fig. 2 Typical microphone arrangement for medium shot. 

times necessary to adjust the null directions of the L and R micro- 
phones on the opposing person. This change in pattern does not 
elirninate pickup into any microphone so adjusted, because the null 
is imperfect and sufficient energy arrives from other directions, thus 
satisfactorily meeting necessary pickup requirements. 

In many instances microphones were grouped in a cluster tighter 
than shown in Fig. 2. This condition carried to its limit, occurs 
where a close-up of two actors is used who are physically separated by 
only three to four feet and farther, they will appear on the screen to 
be eight to ten feet distant from one another, resulting in the crossed- 
over configuration of Fig. 3. Note that the L microphone is placed on 
camera right but is actually picking up the left-hand actor, provided 




of course that it is adjusted to minimize pickup from the right-hand 
performer. The R microphone is, of course, also reversed as indi- 
cated. This, obviously, is one of the most difficult types of pickup, 
as the microphone positions must be carefully chosen, the directivity 
nulls correctly used and, in particularly bad cases, the relative chan- 
nel gain adjusted. It is sometimes wisest to abandon sound place- 
ment under these conditions and use the condition of Fig. 4, which is 
an effective means to maintain stereophonic quality without the fea- 
ture of sound-origin placement. By proper choice of dimensions, un- 
wanted sound placements can be eliminated. Average dimensions 
might be five feet on each side of the triangle and three and one-half 
feet on the base. 

Fig. 3 Typical microphone arrangement for close shot of two actors re- 
quiring "sound magnification" or distance distortion. Note crossed-over 

Manifestly, it is no longer permissible to revert to the demands of 
early sound motion pictures that actors be fixed at specific positions 
for the delivery of dialog. Therefore, motion of microphones is re- 
quired. Combinations of all of the microphone configurations shown 
in the preceding figures have been used in many of the scenes recorded. ' 
The principal problem is one of smooth transition and proper apparent 

Two interesting and useful effects, not possible without stereophonic ; 
methods, have been used. The first creates the illusion of an actor \ 
talking and moving within the set but never being seen while the | 
camera showed a small portion of the set and another player. This ! 
offstage illusion can be created by sound unassisted by the visible j 
actor who, without stereophonic sound, must describe the unseen j 




action by following with his eyes. The second is the ability to make 
sounds from either side apparently very much offstage. 


The equipment used consisted of three essentially identical record- 
ing channels with a common film recording machine, placing three 
200-mil push-pull variable-density tracks on a single 35-mm film. 
The three modulators were arranged in an arc, the two outside optical 
paths brought parallel with two sides of a front surface prism, and the 
center path passed through a hole in the same prism. Separate ob- 
jective lenses were used for each modulator with a single cylindrical 

Fig 4 Microphone arrangement to maintain stereophonic quality but 
without sound placement other than center screen. 

lens near the film for all channels. Pre- and postequalization were 


Monitoring was provided by earphones where each side channel was 
connected to single reproducers on respective ears and the center 
channel was split to both ear receivers in such a way as to supply 
3 decibels less power to each ear than the corresponding side channel, 
the total power from the center channel being the same as the side-, 
channel outputs. Further, the side-to-side cross-feed was not per. 
mitted higher than -25 decibels from the direct source. This 
means, for monitor, was used on all work and found satisfactory. 

The film reproducers used a single exciter lamp, a fixed aperture 
plate with large radius of curvature over which the film passed to 
maintain contact for focus, a single objective lens, three engraved 
slits, and associated photocells arranged in an arc to provide equal 




optical path length. Filtering was similar in principle to the newer 
mechanisms now being supplied and essentially equal in effectiveness. 
The sound reproducers were two-way systems of good character- 
istics. One was located at screen center and the other two placed 
with their axes 2 /s screen width off-center. These dimensions are 
not inviolate. The best arrangement is determined by existing condi- 
tions and desired effect. 


In connection with one two-reel playlet which was used, it was 
necessary to add horse-hoof sounds, footsteps, cup crashes, and shots. 

Fig. 5 Block schematic of re-recording equipment. 

All of these effects, except the cup crashes, were re-recorded from 
stock library material. The equipment arrangements are given by 
Fig. 5. Note that mixers 1 and 2 are conventional 3-channel stereo- 
phonic units. Mixer 3 is a special control designed to transfer the 
single input to any of the three channels or to any two adjacent 
channels, meanwhile maintaining constant total power. With this 
control it is possible to move a single source smoothly back and forth 
to create any desired illusion. 

By the use pf the special control, offstage horses were made to sound 
as though they approached from a distance to the left and came to a 
point just off stage, also gunshots and footsteps were added and 
properly placed. 




The use of re-recording retained the advantages of level smoothing 
and permitted a small amount of placement correction. Dependence 
cannot be placed on re-recording for changes in placement of original 
material because placement, except in certain special cases, is not 
primarily due to intensity differences. This point is developed more 
fully later. 


Large orchestras (90 to 100 pieces) were recorded within a regular 
scoring stage. It was desired to obtain good separation of instru- 
ments and due to the compactness of the arrangement, unidirectional 
microphones were again used, placed in a relatively close group. 


Fig. 6 Typical orchestra and microphone arrangement for a large 
group in a particular scoring stage. 

This also helped to minimize troubles due to room acoustics causing 
false origins. A typical setup is shown in Fig. 6. 

The recordings obtained from one such session were used for play- 
back and the orchestra photographed in a large set. Various cuts and 
angles were used and it was necessary to exercise caution to select 
angles and musical passages which were compatible. In some in- 
stances minor sacrifices to correct sound placement were made to 
provide adequate camera freedom. 

One vocal number was recorded experimentally at the time a regu- 
lar production prescore and vocal-recording session was in progress. 
The vocalist was performing in a small vocal booth with the orchestra 
in the adjoining scoring stage. A separate microphone was provided 
for the vocalist and a stereophonic pickup of the orchestra arranged. 

288 GRIGNON March 

The vocal and center-channel music microphone outputs were mixed 
at the time of the performance, thus obtaining monaural vocal always 
on the center channel and a stereophonic record of the accompanying 
orchestra. This track was later used for playback and the actress 
photographed, but since the vocal existed only on the center track it 
was necessary to frame the action so that the performer was nearly 
always center screen. 

, It was noted that orchestra levels greater than normal could be 
used without destroying the effectiveness of the vocal selection. This 
might be explained as follows: The vocal is always reproduced as a 
single direct source and is audibly compared to a stereophonically 
reproduced accompaniment, thereby increasing the perceptible aural 
differences and subjectively providing greater separation. No work 
of this type was done with both sources recorded by stereophonic 
methods and until this is done and compared to the method reported 
here, no final conclusion can be reached. 


Since this experiment was an integrated project involving all present 
motion picture production methods, demanding close correlation be- 
tween each contributing group, it was possible to evaluate the effect 
that stereophonic-recording application might have on motion picture 
production and presentation in general and various phases in particu- 
lar. Those effects, and other observations based on the work herewith 
reported and which presently seem of the greatest importance, will 
now be discussed. 

From the microphone-pickup work come three cardinal points; 
sound placement matching corresponding picture, smooth placement 
transition of the sound from a moving source, and a third point not 
previously mentioned, that to avoid major changes in quality some 
sound must be picked up in all microphones at all times. 

The requirement of correct sound placement is obvious. It has 
been found that sound-intensity differences do not play the major 
role in determining placement except under unfavorable acoustic con- 
ditio'ns. Those situations in which high-intensity directive reflections 
occur and are then picked up by. a microphone other than the one 
closest to the source create the exceptions. Under such circum- 
stances there exists only a small intensity difference between the 
nearest and- other microphones and an otherwise minor change in 
intensify adjustment can introduce a change in placement. With 


suitable acoustic conditions, intensity differences due to equipment 
maladjustment of 6 to 10 decibels do not destroy localization but 
loudness is of course affected. These observations would indicate 
that the greater contribution to sound placement is caused by phase 
differences which are a complex function of acoustics, frequency, and 
ratio of microphone spacing to frequency. 

Smooth sound transition is necessary, otherwise sudden placement 
jumps occur which are very disturbing to any observer after a short 
acquaintance with stereophonic reproduction. 

The third point concerning quality is related to the inherent im- 
provement in stereophonic over monaural methods. It has been 
demonstrated that a two-channel stereophonic system does not pro- 
vide the quality improvement afforded by a three-channel arrange- 
ment, as might be expected since the former approaches closer to the 
inferior monaural condition. The quality difference between two- 
and three-channel systems is such as to establish the foregoing state- 
ment concerning pickup in all microphones. 

In connection with recording on a production basis two specific 
items of equipment were greatly needed. The sound mixer should 
have a picture monitor displaying the scene the camera is photo- 
graphing. This apparatus will shorten rehearsal time and guarantee 
sound and picture match. Such devices are now available by tele- 
vision technique and are rapidly approaching practicality for motion 
picture use. Second, a better mechanical device than presently used 
microphone booms must be devised. Some combination of mecha- 
nisms which could permit microphone movements with fewer personnel 
is highly desirable, not only for the sake of reducing manpower costs 
but to minimize errors caused by lack of co-ordination. 

Film editing must be considered when using stereophonic sound. 
Directors, photographers, and editors are ever watchful of camera 
angles and actor movement to facilitate smooth cutting, and stereo- 
phonic recording demands the same consideration. Much of this 
problem is automatically solved when the visual action is properly 
done, but consider the effect of an offstage voice from the left when the 
offstage person is shown, in the very next cut, on screen right. When 
cutting pictures editors always strive for action which flows smoothly 
and logically unless spectacular effects are desired and these are then 
introduced deliberately. Exactly the same situation exists with 
stereophonic recording. With proper understanding "jumpy" effects 
can be eliminated or purposely used when apt. At present, picture 

290 GRIGNON March 

editing is hampered very little by sound. It may. very well be that 
some types of cuts or cutting practice cannot be used successfully in 
connection with stereophonic methods, and picture may have to con- 
cede somewhat more importance to sound. 

How will production costs be affected by stereophonic recording? 
Any answer to this question at this stage of the art requires some 
speculation since so many factors contribute to such costs. Expense 
can be greatly influenced by the degree of perfection in the result de- 
sired. Considering some of the more tangible factors, it does not 
appear that production costs need be increased excessively. At no 
time were more than three complete rehearsals required to satisfy 
sound-pickup requirements. Usually one or two sufficed; one of 
these being necessary to observe the action through the camera finder 
to establish relative placements. Rehearsal time could have been cut 
in half had the remote finder mentioned above been available. 
Undoubtedly, with greater experience, the demands of original stereo- 
phonic sound would not be much different than at present. 

Dubbing costs would of course increase, since greater time would 
be involved to match both lip movements and action. However, the 
use of dubbing, while absolutely necessary in some cases, is a dodge 
and is best minimized. 

Demands on set construction are no different for stereophonic than 
for monaural recordings. An acoustically good set monaurally is 
still a good set stereophonically and good acoustics, in general, are 
now sought. Actually, a saving may result in using stereophonic 
methods because of the poor records frequently obtained in portions 
of otherwise acceptable sets which, if sufficiently inferior, are dubbed 
or re-recording time is used up in attempted correction. 

Equipment costs, which are a small percentage of production 
charges, would be raised two to two and one-half times. 

Stereophonic reproduction in the theater naturally will require 
additional equipment. In the event that stereophonic methods are 
applied to motion pictures, some technical and economical method 
must be devised to supply both stereophonically and monaurally re- 
corded film during conversion. It is even likely that such practice 
would continue for some time in order to supply the very small low- 
income theaters and, in some cases, for reduction for 16-mm release. 

1 . Greatly improved sound quality can be obtained by the use of 


stereophonic methods. It is easily demonstrable that recordings 
made in sets which give unnaturally "boomy" or otherwise poor 
results monaurally result in records which more nearly reproduce the 
true conditions in that set when recorded stereophonically. This is 
still true when disregarding subjective sound placement. 

2. Sound placement is affected only to a small degree by individu- 
ual system gain differences indicating that phase and not intensity 
differences play the major role in determining placement. 

3. The three important points of stereophonic pickup are: (1) 
sound placement matching visible or desired implied action; (2) 
smooth sound-placement transition and (3) some pickup in all micro- 
phones at all times. 

4. Many more illusions can be created by sound alone, opening 
new dramatic, effective avenues for motion picture story presentation. 

5. Just as the directions of visual action must be properly done to 
permit cutting, so must stereophonic sound directions be considered. 
Of a similar nature, since it pertains to camera angles and editing, 
prescoring for playback purposes should be planned to match the 
intended action and anticipated cutting. There is evidence that 
present editing practices would need modification. 

6. With sufficient experience and certain desirable auxiliary equip- 
ment, production cost need not be greatly increased. Two of such 
auxiliaries are a picture monitor (remote viewfinder) for the sound 
mixer and more suitable microphone-handling equipment. The de- 
gree of perfection desired would be the largest cost factor. 

7. Re-recording, technically, is no more difficult than at present 
but having introduced one additional degree of freedom, more manipu- 
lation will be required. Many stock library monaural tracks may be 
used, provided equipment is available for controlling placement of the 
desired sound. 

8. Increased effectiveness of stereophonic sound is obtained if used 
with a picture of greater aspect ratio than presently used. Given a 
picture in which the ratio of width to height is approximately 1.75 
instead of the existing 1.33 a somewhat closer approach to the hori- 
zontal angle of human vision is obtained and the relatively greater 
width assists sound placement by simplifying the original pickup and 
giving better coverage hi the theater. 

As with any other subject of similar complexity, no one experiment 
answers all the questions. Much work remains to be done. Repro- 
duction in various kinds of auditoriums has only been superficially 


explored. Some of the questions will only be fully answered by ac- 
tual production experience. 

Contemplation of the results obtained from the described project 
and with a realization of remaining problems, it is concluded that 
stereophonic recording can be used for motion pictures and will pro- 
vide a superior sound presentation which is one step closer to technical 
perfection, and realism on the screen. Unfortunately, stereophonic 
sound cannot be introduced overnight but it can be made available to 
the industry if wanted. 


The welcome assistance of E. I. Sponable of Twentieth Century- 
Fox Films and K. F. Morgan of Western Electric, as. well as several 
others, is acknowledged. 


(1) Bell Telephone System, Monograph B-784, "Auditory Perspective," Sym- 
posium of six papers, 1934. 


QUESTION: Have you ever tried to balance two microphones in a derby hat 
and use this as a pickup for sound? We have found in this way, we can get ex- 
cellent reproduction, even better than with three microphones placed at such long 
distances. I think the reason for this is that, as you see in your picture, you will 
have only time differences between the microphone. 

MR. L. D. GRIGNON: No, we have riot tried that particular combination. We 
assumed that a good starting point wonld be based on the extensive previous 
work referred to in a Bell System Monograph. It was our job to try to adapt 
stereophonic as it was then known to the motion picture problem. 

QUESTION: During the war we were able to do much work on stereophonic 
sound in Holland, and I think we found a better principle to start from than you 
did here in America. 

DR. J. G. FRAYNE: If you make an error in the original, how far can it be 
corrected in re-recording process? How can you switch people from right to left 
if you do not get the original track straight? 

MR. GRIGNON: Possible corrections depend upon the degree of the error. If 
it is a minor error, you can push the intensity difference enough to make some 
correction, but if it is a major error, nothing can be done. 

CHAIRMAN C. R. DAILY: With regard to release track in stereophonic, does one 
require the same quality of reproduction from each individual track that we now 
feel is desirable from single theater tracks? 

TVlR. GRIGNON: Probably not, although let us put it this way: If a major 
change of this kind were to be made, certainly we should take advantage of the 
opportunity to try to increase the fidelity of recording and reproduction. These 
three channels were somewhat better in quality arrangement than are commonly 
used. With three channels, let us say, of the ordinary type now used, the subjec- 
tive quality in reproduction would still be much better than we now have. On 
that basis, we could take less and still come out equally well. 

Single -Element 
Unidirectional Microphone* 


Summary A single-element unidirectional microphone has been de- 
veloped for use in sound motion picture recording with the following charac- 
teristics: single-ribbon type; the back of the ribbon is coupled to a damped 
folded pipe and an acoustical impedance in the form of an aperture; improved 
cardioid directional pattern; greater output; reduced weight; and reduced 
wind-noise response. 


FROM THE INCEPTION of sound reproduction it was apparent to 
those associated with the problems of sound pickup that some 
form of directivity was desirable in the sound-collecting system to 
improve the ratio of direct to reflected sounds and thereby improve 
the reverberation characteristics and otherwise discriminate against 
undesirable sounds. Horns and reflectors were used for the early 
directional sound-collecting systems. As the fidelity of reproducing 
systems was improved, it became apparent that considerable distortion 
in the form of frequency discrimination in both the direct and re- 
flected sounds was introduced because the directional characteristics 
of the horn and reflector systems varied with frequency. About two 
decades ago the velocity directional microphone 1 was developed 
which exhibited uniform directional characteristics over the entire 
audible spectrum. This microphone established the usefulness and 
superiority of a sound-collecting system with uniform directional 

The conventional velocity microphone consists of a single mass-con- 
trolled ribbon with both sides freely accessible to the medium. The 
many desirable performance characteristics exhibited by this micro- 
phone may be attributed to the obvious simplicity of the vibrating 
system. The constants of the system may be chosen so that response 
and directional characteristics will be uniform over the entire audible- 

* Presented May 17, 1948, at the SMPE Convention in Santa Monica. 





frequency range. The nonlinear distortion for the intensity range of 
the human ear is a small fraction of a per cent. The light-mass-con- 
trolled system insures good transient response. 

The polar-directional characteristic of the velocity microphone is 
bidirectional. For certain sound-pickup problems, unidirectional 
characteristics are more desirable. Accordingly, shortly after the 






Fig. 1 Electrical system, vibrating system, 
and acoustical network of the single-element 
unidirectional microphone. 

development of the velocity microphone, a unidirectional micro- 
phone 2 - 3 with uniform directional and frequency response over a wide 
frequency range was developed. The conventional unidirectional 
microphone consists of the combination of a pressure element and a 
velocity element. The original unidirectional microphone employed 
ribbon elements for both the velocity and pressure elements. Em- 
ploying ribbon elements makes it possible to maintain uniform phase 


relations between the velocity and pressure elements without resorting 
to correcting networks. The acoustic fidelity of the unidirectional 
microphone is essentially the same as that of the velocity microphone. 
In view of the importance of directional microphones and the high 
fidelity of ribbon transducers, work has been continued on these sys- 
tems with the object of increasing the scope and simplifying the 
vibrating system. In particular, a microphone has been developed for 
sound motion picture recording with the following characteristics: 

Fig. 2 Assembled single-element unidirectional microphone (production 

model) . 

(1) Single-ribbon type. 

(2) Improved cardioid directional pattern. 

(a) In the high-frequency range due to a smaller vibrating 

system and suitable magnetic structure. 

(b) In the low-frequency range due to the use of an addi- 

tional acoustical element. 

(3) Greater output. 

(4) Reduced weight. 

It is the purpose of this paper to describe the single-element unidirec- 
tional microphone with the characteristics listed above. 


The electrical system, the acoustical system, and the acoustical 
network of the single-element unidirectional microphone are shown in 
Fig. 1. Referring to the acoustical network, it will be seen that this 
is of the bridge type. 




The phase relations between the two actuating pressures varies with 
the direction of the incident sound. For example, for sound incident 
upon the back of the microphone, the time required for the sound to 
pass through the hole in the pipe to the back of the ribbon is the same 
as the time required for the sound to pass around the magnet structure 
to the front of the ribbon. Under these conditions, the same pressure 
with the same phase is exerted on the front and back of the ribbon, and 
as a result the ribbon does not move. For sound incident upon the 
front of the microphone, the sound which appears at the back suffers 
a delay caused by the path around the magnet structure and a delay 

Fig. 3 Front view with the screen removed. 

through the hole in the pipe. Therefore, there is a relatively large 
phase angle between the sound pressures on the two sides of the ribbon 
which means that the ribbon will move because of the difference in 
pressure due to the difference in phase. For the particular phase rela- 
tions that exist between these pressures for different angles of inci- 
dence, it is possible to choose the constants of the system so that a 
cardioid characteristic will be obtained. 

It was found that the deviation from a cardioid pattern at the lower- 
frequencies was caused by the unbalance produced by the acoustical 
resistance in the branch z Ai . By introducing a corresponding acous- 
tical resistance in the branch z A 2, the* balance can be restored. This 


state of affairs can be deduced from the acoustical network and the 
following theoretical considerations. 

In the acoustical circuit the elements are as follows : 

In branch Z A I 

Ms = inertance of the slit between the ribbon and pole 

r A s = acoustical resistance of the slit between the ribbon and pole piece 

M R = inertance of the ribbon 

FAR = acoustical resistance of the ribbon 

CAR = acoustical capacitance of the ribbon 

MA inertance of the air load upon the ribbon 

r AA = acoustical resistance of the air load upon the ribbon 

ZAE = acoustical impedance due to the electrical circuit. 

It is given by 

ZAE = 9 I ^ (!) 


# = flux density in the air gap 

I = length of the ribbon 

ZEI = electrical external load on the ribbon (see electrical circuit, Fi?. 1) 

r E i = electrical resistance of the ribbon. 

z A i is the acoustical impedance of the branch composed of the above 
In branch z A2 

MI = inertance of the aperture 

M 3 = inertance of the screen covering the hole 

r A3 = acoustical resistance of the screen covering the hole 

A/4 = inertance of the air load upon the screen and aperture 

r A4 = acoustical resistance of the air load upon the screen and aperture. 

z A 2 is the acoustical impedance of the branch composed of the above 
In branch Z AS 

ZAS = TAP = acoustical resistance of the acoustically damped pipe. 

The sound pressure acting on the open side of the ribbon may be 

Pi = POI CJ<<*t+*l> (2) 

Poi = amplitude of the pressure 

CO =27T/ 

/ = frequency 

t = time 

<t>i = phase angle with respect to a reference point. 


The sound pressure acting on the aperture in the labyrinth connector 
may be written 


Poa = amplitude of the pressure 

<fc = phase angle with respect to a reference point. 

The reference point for the phase may be changed so that 

Pi = Poi ^'O") (4) 

and p 2 = po,, eJ'M+03). (5) 

The phase angle </> 3 is a function of the angle of the incident sound as 

<fo = cos 6 (6) 


= angle between the normal to the surface of the ribbon and the direction of 

the incident sound, and 
= phase angle determined by the dimensions and geometry of the ribbon and 

the structure surrounding the ribbon. 

The volume current of the ribbon due to the pressure pi is 
. . X = Pifciz + Z AS ) 

*Al*A* + Z A iZ AS + Z A2 Z A3 ' 

The volume current of the ribbon due to the pressure p z is 


z A iz A z -n 2412.43 T z A zz A3 



The resultant volume current X R of the ribbon is the difference be- 
tween (7) and (8). 

X R = X, - X 2 . (9) 

The internal voltage generated by the motion of the ribbon is given 



?= flux density in the air gap 
= length of the ribbon 
A = area of the ribbon 
X R volume current of the ribbon given by (9). 




Fig. 4 Rear view with the screen removed. 

Fig. 5 Exploded view showing bottom cover, cover plate, folded and 
damped pipe, screen^ magnetic and vibrating structure, transformer case, 
and transformer. 




A consideration of the above system shows that it is theoretically 
possible to obtain a true cardioid characteristic by a proper choice of 
constants. However, in a practical microphone certain incompatible 
factors appear to make this impossible. For example, high sensitivity 
requires a large magnetic structure. On the other hand, the simplest 
way to obtain a true cardioid characteristic is to use a vibrating sys- 
tem and surrounding structure which is small compared to the wave- 
length. The sensitivity of a microphone with such relatively small 
dimensions would be too low. However, it is possible to reduce the 
phase effects in a relatively large structure by a suitable design. In 

ZOO 400 1000 2000 


7000 10000 15000 

Fig. 6 Response-frequency characteristics of the single-element uni- 
directional microphone for the incident angles of 0, 45, 90, 135, and 180 

order to determine the optimum compromise between sensitivity and 
directivity, a study was made of the magnetic and acoustic character- 
istics of twenty structures. The selected structure for this micro- 
phone will be described in the next section. 


Photographs of the final model of the single-element unidirectional 
microphone are shown in Figs. 2 to 5, inclusive. Fig. 2 shows the 
completed microphone with the wind screen in place. In Figs. 3 and 
4 the wind screen has been removed. Fig. 3 shows a front view, 
depicting the magnetic structure and ribbon. Fig. 4 is a rear view, 
depicting the magnetic structure and labyrinth connector. The 






180 7o 

7000 CYCLES 10000 CYCLES 

Fig. 7 Polar-directional patterns of the single-element unidirectional 
microphone for the frequencies of 50, 200, 1000, 7000, and 10,000 cycles per 


acoustical resistance, in the form of a screen over the hole in the laby- 
rinth connector, can also be seen in Fig. 4. Fig. 5 is an exploded view, 
showing all the component parts of the microphone. 


The measured response-frequency characteristics and directional 
characteristics of the single-element unidirectional microphone are 
shown in Figs. 6 and 7. It will be seen that the directional pattern is 
practically independent of the frequency of the incident sound over 
the frequency range from 50 to 8000 cycles. Furthermore, a high 
order of cancellation for sounds arriving from the rear is obtained 
from 50 to 14,000 cycles. The high order of cancellation at the low 
frequencies is due to the addition of the acoustical-resistance element. 
The higher order of cancellation at the high frequencies is caused by 
the design of magnetic structure which exhibits uniform sound dif- 
fraction over the frequency range up to 8000 cycles. The uniform 
directivity has not been obtained at the expense of sensitivity because 
the open-circuit output for an impedance of 250 ohms is 260 millivolts 
per dyne per square centimeter at 1000 cycles. The light weight of 
two pounds is quite low for this order of sensitivity and frequency 


(1) H. F. Olsen, "Mass controlled electrodynamic microphone: The ribbon 
microphone," /. Acous. Soc. Amer., vol. 3, pp. 56-68; July, 1931. 

(2) H. F. Olson, "A unidirectional ribbon microphone," /. Acous. Soc. Amer., 
vol. 3, pp. 315-316; January, J932. 

(3) H. F. Olson and J. Weinberger, U. S. Patent, Reissue 19115, 1934. 

16-Mm Film Phonograph 
for Professional Use* 



Summary Superior performance of new 16-mm film phonograph designed 
for the motion picture industry permits its use as a standard to determine the 
sound quality of 16 mm films and to check performance of 16-mm projectors. 

Additional features include a self-contained preamplifier, rewind at acceler- 
ated speed, dependable operation, compactness, accessibility for servicing, 
and attractive styling. 

SINCE THE INCEPTION of sound on film, 35-mm film has been the 
accepted standard of the motion picture industry. For several 
reasons, 16-mm film has not been used where the best in picture and 
sound were required, but 16-mm film production is on the increase 
and there is a tendency in this to judge the quality of 16-mm sound in 
direct comparison with that obtained with 35-mm film. With the 
increase in the number of 16-mm productions has come an increase in 
the use of 16-mm equipment exclusively in these productions. Much 
of the original recording is being done on 16-mm film. Therefore, not 
only is the quality of the performance of the 16-mm film recorder 
important to the achievement of high-quality sound print, but also 
the quality of performance of the 16-mm film phonographs used in 
the re-recording of the productions and in determining the quality 
of the original recording. A paper by Collins 1 described the con- 
struction and performance of a new 16-mm film recorder. The sound 
quality obtainable with it compares favorably with that obtainable 
using 35-mm recorders of modern design. The PB-176, 16-mm film 
phonograph (Fig. 1) has been designed as a companion unit using all 
the applicable features of the PR-32, 16-mm recorder described in 
the aforementioned paper. 

The principal features which the two units have in common are: 
(a) A base assembly (Fig. 2) containing the receptacles for electri- 
cal connections, lamp rheostat, and cast-in recesses that form the 
handles for carrying. 

* Presented May 20, 1948, at the SMPE Convention in Santa Monica. 





Fig. 1 PB-176 16-mm film phonograph. 

(b) Covers for the optical, motor, and gear compartments. 

(c) A head assembly (Fig. 3) containing all of the gearing and film- 
handling equipment. 

(d) A driving motor of synchronous or interlock type for either 60- 
or 50-cycle power. 

(e) A control-panel assembly containing lamp-rheostat adjust- 
ment, lamp ammeter, and switches for motor and lamp. 

(f) Enclosed-type beltless take-up and rewind assembly. 

Fig. 2 Rear view. 


The film phonograph differs from the recorder mainly in that it 
contains a reproducing optical system plus a phototube preamplifier 
in place of the recording optical system (Fig. 4). 

Principal mechanical features of the PB-176 film phonograph which 
deserve special notice include : 

(a) Film motion having total flutter less than 0.1 per cent. 

(b) Lateral film weave of less than 0.001 inch. 

(c) Machine noise of 53 decibels below 100 per cent modulated 

Fig. 3 Rear view, covers removed. 

(d) Film compartment of enclosed type having a wide-opening 
door for easy access to sprockets, rollers, and that portion of the opti- 
cal system installed within this compartment. A wide slot on the 
upper side of the compartment simplifies threading to and from the 
take-up and rewind reels mounted on reel spindles above the 
compartment. Guide rollers are provided to direct the film into the 

(e) Take-up of enclosed positive-drive type capable of handling 
800-foot reels. (For those special applications where 1600- or 2000- 
foot reels must be used, a take-up drive assembly will be available for 
reels of that size.) 

(f) Rewind which is enclosed with take-up mechanism and which 
automatically utilizes a portion of take-up gearing driven in the 




reverse direction. To rewind, the film is threaded directly from the 
right-hand reel to the left-hand reel and the motor is then run in 
reverse rotation by throwing the reversing switch mounted on the con- 
trol panel to the right of the film compartment. In the synchronous- 
motor-driven units the reversing switch reverses two motor leads 
thus reversing motor phasing and causing the motor rotation to be 
be reversed. In the Selsyn-motor-driven units, even though the 

Fig. 4 Optical system and phototube amplifier. 

motor circuit is more complex than that of the synchronous motor, 
operation of a single switch on the control panel also will cause the 
motor rotation to be reversed. This switch is a four^pole double- 
throw type with a positive neutral or off position. In the up or nor- 
mal" running position, the motor is connected to the Selsyn distribu- 
tor. In the down or reverse position of the switch, the stator of the 
motor is connected to a separate three-phase power supply whose 
phasing is the reverse of that for Selsyn operation and the rotor leads 
are short-circuited together through a relay mounted in the base 
of the film phonograph. This permits the motor to operate as an 


induction-type motor at nearly one and one-half times normal speed. 
The rewind time for a film phonograph equipped with a 60-cycle 
Selsyn motor is 2*/2 minutes for 800 feet of film. 

The new improved reproducing optical system is of advanced de- 
sign in keeping with the remainder of the film phonograph. This unit 
is mounted to the side of the film-compartment portion of the main 
case. Mounting is in such a manner as to provide easy lateral ad- 
justment of the scanning-light beam by means of an adjusting screw. 
This assembly includes: 

(a) The slit and objective-lens assembly which is of comparable 
quality to those used in theater-type soundheads. Azimuth adjust- 
ment is provided by two opposing setscrews which rotate only the 
slit-condenser and slit-plate assembly. Focusing of the objective is 
accomplished by turning the lens mounted in a keyed barrel anol an 
instantaneous change-over focus adjustment is provided, by rotating 
a calibrated ring equipped with a finger pin, for films having their 
emulsion either toward or away from the scanning-light source. 

(b) An exciter lamp which is of the standard prefocus type with 
curved filament designed to operate at 10 volts, 5 amperes. Con- 
venient vertical adjustment of the lamp filament is provided. A 
quick release lever on the lamp socket permits easy removal of a de- 
fective lamp. 

(c) The phototube, RCA Type 927 Radiotron, is mounted on the 
optical bracket and extends from it into the film compartment above 
the sound drum. The phototube is readily accessible from the film 
compartment and may be removed after removing its slip-on-type 
light shield. The scanning-light beam, after passing through the 
film, is efficiently directed to the phototube by a single-unit prism and 
lens assembly mounted adjacent to the sound drum. 

A compact phototube amplifier is provided as an integral part of the 
film phonograph. It is mounted in the optical compartment behind 
the optical system and is accessible for servicing or removal from the 
film phonograph by removing the optical compartment cover which 
is held in place by two thumbscrews. In order to obtain maximum 
efficiency, the phototube mentioned above is directly connected, by a 
short low-capacitance cable, through a high-impedance circuit to the 
grid of the first preamplifier stage. Electrical connections to the 
phototube amplifier are made at conveniently located externally 
mounted terminal boards. The preamplifier contains a power switch 

308 HlTTLE 

and gain control. Electrical characteritics of the phototube amplifie r 
are as follows: 

(a) Output impedance: 150/250 ohms (balanced or unbalanced) 

(b) Filament: Adjustable from 6.3 to 14 volts direct current 

(c) Plate: 250 volts direct current 

(d) Output level: 2 dbm* for normal operation and +8 dbm max- 


(e) Noise level : 53 decibels below 100 per cent modulation 

(f ) Frequency response: Flat characteristic within == 1 decibel from 30 to 6000 

cycles using Vz-mil optical slit and flat film 

(g) Polarizing voltage: 80 volts direct current 

(h) Tube complement: Two RCA 1620 Radiotrons 

(i) Distortion: Less than V 2 per cent from 50 to 7500 cycles at +4 

dbm output 
(j) Gain control: 7-decibel range 

* Decibels with respect to 0.001 watt. 

The many self-contained features such as a conveniently located 
exciter lamp and motor controls consisting of an ammeter, rheostat, 
and switches, plus phototube amplifier; the selectivity offered of 
take-up and rewind sizes, the compactness of the design, the pleasing 
styling, and the convenience and dependability of operation make the 
PB-176, 16-mm film phonograph a distinctive unit physically. Its 
equally distinctive performance characteristics further enhance its 
desirability by those engaged in 16-mm work for re-recording activity 

or for carefully checking recorded material. 


Recognition is given to J. L. Pettus, L. T. Sachtleben, K. Singer, 
and E. P. Ancona for their contributions to the design and testing of 
this equipment. 


(1) M. E. Collins, "Lightweight recorders for 35- and 16-mm film," /. Soc. 
Mot. PicL Eng., vol. 49, pp. 415-425; November, 1947. 


MR. HARRY ERICKSON : Would you please repeat the distortion effect figure? 
MR. CARL E. HITTLE: The figure is less than one-half per cent, from 50 to 7500 
cycles per second at plus four dbm output. 

Audio-Frequency Oscillator 
for Calibrating 
Flutter-Measuring Equipment* 



Summary This paper describes an electronically frequency-modulated 
oscillator of the resistance-capacitance type which has been developed to 
facilitate rapid and accurate calibration and testing of flutter-measuring 
equipment; specifically, of the portable flutter indicator used in theater 
service work. It supersedes mechanically driven capacitor devices for pro- 
ducing the frequency-modulated 3000-cycle signal used for flutter-indicator 
calibration. It possesses the advantages of better frequency stability, purer 
wave form, and absence of spurious output impulses. 

FOR MANY YEARS one of the major problems inherent in all types 
of sound recording and reproducing methods has been that of 
attaining constant speed of the record medium, whether film, flexible 
tape, wire, or disk, at the recording or scanning point. The extent of 
this problem has been shown by the many articles which have been 
published in the JOURNAL of the Society of Motion Picture Engineers, 
the Proceedings of The Institute of Radio Engineers, the Journal 
of the Acoustical Society of America, and others, concerning the work 
done in the effort to minimize speed variations. 

In many present-day recording and reproducing equipments such 
undesirable speed variations are reduced below the level of audible 
disturbance. The attainment of this desirable level in operation 
requires care, both in making the original installation and in the 
maintenance of the equipment. To ensure satisfactory maintenance, 
various instruments and test methods have been develped to assist 
the field engineer. Among the first of these was the portable flutter 
indicator, developed by the Radio Corporation of America for meas- 
uring flutter in theater sound equipment and made available to its 
theater service field men in 1938. 

* Presented May 20, 1948, at the SMPE Convention in Santa Monica. 



Previously, the calibration of these flutter-indicator instruments 
was accomplished by the use of a frequency-modulated oscillator 
whose 3000-cycle fundamental frequency could be varied (by means of 
a rotating motor-driven capacitor) plus or minus a maximum of 2 
per cent in selectable increments of 0.1 per cent. The design of 
this motor-driven capacitor presented some interesting problems. 
Previous experience with mechanically operated, frequency-modula- 
tion devices indicated that connections to the rotating plates would 
be troublesome, and therefore, a symmetrical double rotor and stator, 
with the circuit connections made to the two stators, would be neces- 
sary. While the symmetrical double rotor aided in obtaining me- 
chanical balance of the rotating parts, the shape of the plates required 
to obtain symmetrical sine-wave modulation when paralleled with the 
remaining circuit capacitance was rather unique. Further, as the 
required amount of frequency modulation was, of necessity, a variable 
ranging from to 2 per cent, the addition of the required series and/or 
shunt capacitors to the rotating capacitor circuit changed the sym- 
metry of the modulation. Mechanical vibration of the moving parts, 
together with hysteresis of the iron core of the oscillator coil, in- 
troduced unwanted and random modulation. Any looseness or wear 
in the bearings of the rotating capacitor also contributed to instability 
of the output. As the rate of frequency change was controlled solely 
by the shape of the rotating capacitor plates, various modulation 
patterns could be obtained only by designing and making a special 
capacitor for each pattern desired. 

After considering all of the deficiencies of the motor-driven capac- 
itor oscillator, it was felt that an electronic type of oscillator which 
would be free from the inherent limitations of the mechanically vari- 
able type should be designed. Similar oscillators have been de- 
veloped in recent years for various applications, notably the RCA 
facsimile transmitter. The circuit shown in Fig. 1 is the outcome of 
considerable research work and study applied to the basic circuit. 
Scrutiny of this circuit will reveal that a resistance-capacitance 
phase-shift oscillator is the heart of the device. 

To vary the frequency of this oscillator at the desired rate, one of 
the resistances in the network is composed partly of a physical resistor 
and partly of the plate resistances of a push-pull pair of triodes. 
Control voltage applied to the grid of the lower triode section results 
in practically instantaneous change of plate resistances and, there- 
fore, of the frequency of the oscillator. The amplitude of the 




oscillation remains substantially unchanged through considerable 
variation in frequency, and therefore, the output of the oscillator is 
free from unwanted amplitude modulation. 

The 2000-ohm control rheostat connected in parallel with the 200- 
ohm cathode resistor of V-3 (which together furnish grid bias for the 
control tube), is provided to set the oscillator frequency initially 
to 3000 cycles per second. Plate voltage to the circuit is stabilized 
by the voltage regulator tube V-5. The frequency-modulated out- 
put of the oscillator is fed into an audio amplifier to obtain the re- 

Fig. 1 Frequency-modulated audio oscillator. 

quired amount of driving power for proper indication of the flutter 
bridge being calibrated. 

The input to the grid of the control triode may be of any frequency 
from zero to several hundred cycles per second and of any wave shape 
from extremely peaked through sine-wave to rectangular. For 
sine-wave frequency modulation, an audio-frequency oscillator of the 
familiar beat-frequency or resistance-capacitance type may be used. 
For other types of modulation, such as square wave, appropriate 
generators are available. 

To obtain the precise amount of frequency modulation desired 
within the range of to 2 per cent, a thermocouple-type voltmeter 


and an attenuator are built into the instrument. The voltmeter 
indicates the root-mean-square voltage applied to the voltage- 
divider attenuator. The attenuator taps correspond to the per- 
centage points for which the flutter-bridge scale is marked. 

The use of a thermocouple voltmeter insures that regardless of the 
wave shape of the modulating voltage used, the percentage modula- 
tion of the oscillator will be correct root-mean-square values as 
stipulated elsewhere. 1 

A further advantage of the use of a thermal voltmeter is the fact 
that at modulation frequencies approaching zero the thermal inertia 
of the couple prevents unsteady indication of the voltmeter pointer. 
As this instrument is a laboratory device, the extra sturdiness of 
other types of voltmeters is considered unnecessary. 

Since the thermocouple voltmeter indicates root-mean-square 
values directly, standardization of the instrument is easily made 
with direct current. This is done by applying direct and reversed 
direct current to the grid of the control tube, and measuring the 
voltage required to give the required frequency deviation as indicated 
on a Conn stroboscope. The required attenuator resistors are cal- 
culated and the input voltmeter scale marked at one scale point 
corresponding to the required attenuator input voltage. The ac- 
curacy of the attenuator taps is then checked by comparison with the 
Conn stroboscope, and by any necessary readjustments made to the 
attenuator resistors. 

This instrument is free from amplitude-modulation effects, gives 
excellent frequency stability, good output wave form, long-time con- 
stancy of calibration, and provides rapid checks of flutter-bridge 
scale markings. It may be modulated readily at any low audio 
frequency, and with any wave shape desired by using an external 
modulation generator. Its use has enabled extensive investigation 
.of flutter-indicator characteristics, which will in turn permit more 
accurate field tests of sound-reproducing equipment. 


(t)~ "Proposed standard specifications for flutter or wave as related to sound 
records," /. Soc. Mot. Pict. Eng., vol. 49, pp. 147-160; August, 1947, section 4.1. 

Silent Playback and 
Public-Address System* 


Summary The silent playback and public-address system provides a 
unique time-saving aid in photographing and recording dialog sequences hav- 
ing musical backgrounds. By means of a music- or speech-modulated 100- 
kilocycle transmitter, a supersonic-frequency magnetic field is generated 
throughout the scene area. The magnetic field includes secondary currents 
in concealed loops worn by the actors. These signals are demodulated in 
miniature batteryless units and energize small hearing-aid earphones. The 
earphones may be worn without concealment on distant and medium scenes. 
On close-ups the earphones are concealed and the sound is coupled to the ear 
through small plastic tubing camouflaged by make-up. Several types of 
loops, demodulators, and earphone units meet the varying needs of the scenes 
to be photographed and recorded. The stage microphone and sound-on- film 
recording channel is free from 100-kilocycle interference. 

IT HAS BEEN generally recognized that the cost and complexity of 
photographing and recording many motion picture scenes would 
be reduced if the actors could be cued, or directed, or could be en- 
abled to hear a scene's music, by means of earphone units, small 
enough to be disguised by make-up, energized by miniature radio 
receivers concealed in the clothing. The recording-system micro- 
phone would hear only the desired sounds while the actors would be 
guided through complex scenes by the director, this, in a manner 
reminiscent of the silent motion picture method of direction. In the 
case of dialog lines read through a musical background, certain actors 
could speak their lines to the recording-system microphone while 
other actors might dance through the scene to music heard by way of 
disguised earphones and receivers. Or an actor wearing the mini- 
ature equipment could sing a song to the silent accompaniment of a 
temporary piano recording which later would be replaced by a re- 
cording of a full orchestra. In each of the previous cases, the proper 
balance between the music and speech would be restored in the re- 
recording operations. 

* Presented May 18, 1948, at the SMPE Convention in Santa Monica. 





An assignment to design and construct this type of equipment was 
given the authors. Considering only the studio-production require- 
ments, it was decided that while making the receivers as small as pos- 
sible, also to make them batteryless, expendable in case of trouble, and 
to build a transmitter powerful enough to cause the receiver to oper- 
ate with good volume. Receivers were constructed using germanium- 
crystal detectors and the miniature hearing-aid-type capacitors and 

Fig. 1 Loop receiver, earphone, and coupling unit. 

connectors. The receivers energize hearing-aid-type phones coupled 
to the auditory canal with either commercial-type adaptors or es- 
pecially designed adaptors coupled by plastic tubing. 

However, sufficient transmitted energy to insure an adequate out- 
put of the crystal-detector receivers, would cause a large radio- 
frequency signal to be radiated. To avoid this, a novel scheme is 
used. Using a speech-modulated, 100-kilocycle-per-second trans- 
mitter, a current is sent through a closed single turn of wire surround- 
ing the set. A high-frequency magnetic field is thereby generated 
throughout the set area. This field induces secondary currents 
in multiturn loops worn by the actors. These loops of from 6 to 10 


inches in diameter are worn in the hairdress or under the clothing of 
the actors. In effect, the single turn of wire about the set is the 
primary of an air-core transformer while each loop is a separate 
secondary. The induced secondary currents, demodulated in the 
germanium-crystal-detector receiver, energize the earphone units. 

Fig. 1 shows the completed receiver with its protective covering. 
The miniature Western Electric Type 400-C germanium rectifier is 
used for detection. The tuning capacitor, rectifier, load-resistor, 
and blocking capacitor are built into a small terminal cord to which 
the loop is fastened and connected. The receiver circuit is con- 
ventional except that no by-pass capacitor is used across the ear- 
phones. The advantage in quality with a by-pass capacitor was small 
and, in the interests of size, it was removed. The loop has an in- 
ductance of 1.25 millihenries. A 0.002-rnicrofarad capacitor in 
parallel with the loop tunes the circuit to the transmitter frequency. 
After each unit is tested, the loop is covered with split rubber tubing 
that has been recemented, and the equipment cord is covered with an 
air-drying rubber compound. The hearing-aid phone unit is of the 
crystal type and is connected to the loop receiver by flexible leads 
equipped with miniature connectors. 

Connection to the auditory canal of the ear is made in several 
ways depending upon the conditions of use. If the actor is in a 
close-up camera angle a small plastic elbow fitting is adapted to the 
canal by means of various sized rubber inserts. The inserts are avail- 
able commercially. The elbow terminates a thin walled plastic tube 
that is coupled to the phone unit. The tubing, properly coupled, has 
very little voice-frequency attenuation, and lengths up to 30 inches 
may be used. 

Upon occasion, especially fitted earmolds have been made for a 
particular actor or actress by a hearing-aid-equipment laboratory. 
The earmolds are designed to insure the best possible degree of in- 
visibility. They are comfortable and easily fitted to the ear of the 
user. Plastic tubing couples the earmold to the phone unit. When 
the actors who are to receive directional or music cues are at some 
distance from the camera, they may be equipped with commercial 
adaptors. The adaptors are available in a range of sizes for either 
ear. These are of the same type worn for test purposes by hard- 
of -hearing persons while being fitted with hearing-aid equipment. 

A woman's formal attire does not help with the disguise of the 
equipment. To solve this problem a smaller loop receiver was 


designed for use in the hairdress (Fig. 2) . This loop receiver is almost 
as sensitive as the larger receivers. A hairdresser assists in con- 
cealing the elements of the system. Splicing equipment is available 
to enable the fitter of the ear-coupling equipment to shorten or 
lengthen the plastic tubing to the most desirable length. 

The 100-kilocycle-per-second transmitter, Fig. 3, does not differ in 
many ways from conventional radio-frequency units. The induct- 
ances, tuning capacitors, and by-pass capacitors are larger than those 
used in broadcast-frequency equipment. The plate circuit of the 100- 
Jdlocycle-per-second power amplifier is tuned by a variable induct- 

Fig. 2 Standard loop receiver and small hairdress loop receiver. 

.ance and a group of fixed capacitors. A tuning capacitor for this 
frequency, and for the voltage used, would be of an awkward size; 
therefore, the capacitor values are selected by a tap switch. Fine 
tuning is accomplished by a variometer-type rotor in the electrical 
center of the inductance coil. 

The 100-kilocycle-per-second oscillator is crystal-controlled and 
uses Type 6F6 tubes in push-pull. The oscillator drives a buffer 
:stage that uses Type 807 tubes push-pull pentode-connected. Fre- 
quency multiplication is not used, all amplification of the high fre- 
quency being at the frequency of 100 kilocycles per second. The 
power stage is a Class C stage using Type 812 tubes. A step-down 
close-coupled secondary winding, connected to the loop surrounding 
the set area, is wound over the power-amplifier plate-inductance coil. 




The dimensions and resistance of the loop change the effective in- 
ductance of the tuning circuit, making it necessary to retune for dif- 
ferent sized loops. This is the transmitter's only critical adjustment. 
The speech amplifier is conventional in most respects. Dual 
inputs, one for the microphone and one bridging the playback circuit, 
are provided. Provision is made, by means of a relay operated from 
the "push-to-talk" push button on the microphone, not only to make 
operative the microphone circuit, but also to reduce the playback 
volume when the microphone is used. This prevents the masking; 
of the director's voice by the playback music. The amount of this- 
change is adjustable. A switch changes the microphone amplifier 
stage to a 400-cycle-per-second resistance-capacitance oscillator. 
This tone is used to check the strength and distribution of the set- 

100 K.C. 



i. ten 

Fig. 3 100-kilocycle-per-second transmitter block schematic. 

area magnetic field and the sensitivity of the receivers. The maxi- 
mum signal level heard in the phones has been more than sufficient 
to provide the wearer of the equipment a loud, clear signal under all 
conditions. The field is uniform enough that no volume control 
is necessary on the receivers; the audio level being adjusted in the 
speech amplifier to a level suitable to the actors. 

The modulator is a conventional Class B amplifier using Type 811 
tubes, coupled to a modulation transformer which plate-modulates- 
the Class C high-frequency power stage. The modulator plate- 
current meter is used as a speech-volume indicator. 

Although the sound-recording microphone is used in the field,, 
no interference from the "silent" unit has been heard in the sound- 
recording system, except in instances of cable-shield trouble, in which 
cases the device acts as a tool to locate such difficulties. 

The transmitter is powered from 115 volts alternating current. 
The power-switching circuit includes a time-delay relay to prevent 
application of the plate voltages until filaments are heated ; an over- 


load relay to prevent damage to the high-power tubes in the case of 
overmodulation or mistiming; a series ballast resistor to reduce 
power to the high-power tubes during the tuning of the transmitter; 
and door-interlocks to turn off power when high-voltage-area pro- 
tective covers are removed. The steps of power-switching are in- 
dicated by pilot lamps. The exciter, power amplifier, and modulator 
stages are equipped with plate-current meters, and a thermocouple 
ammeter reads the loop current. The transmitter unit (Fig. 4), as- 
sembled from the experimental equipment, has been built into a 
portable dolly. Storage facilities for the loop receivers, phones, and 

Fig. 4 Transmitter and auxiliary equipment 

phone attachments, and for cables and spare parts, have been built 
into the dolly. 

From a production standpoint, this silent playback and public- 
address system has proved itself to be a time-saving tool. The 
recording of certain musical sequences has been remarkably simpli- 
field. In one dancing sequence, twenty loop receivers were in opera- 
tion at one time; forty are available. It has also found service 
in the recording of tap-dance sound effects where the usual earphones 
and their connecting cords have limited the dancers' movements. 
In its first four months of operation, the equipment was used on six 
different pictures with gratifying compliments from the production 
office. A second transmitter has been built to meet increased de- 
mands for this type of service. 



QUESTION : What is the power output of the transmitter? 

MR. BRUCE H. DENNEY: It is a 200-watt transmitter; I believe the actual 
power output is probably in the neighborhood of 50 watts, due to the inefficiency 
of the final stage coupling. It was a bit of a problem. We tried many different 
methods to gain the greatest efficiency. 

MR. GEORGE LEWIN : Are the loops around the set in a vertical plane? 

MR. DENNEY: Horizontal plane, either on parallels ten feet above the floor or 
just laid in a carpet around the stage or directly outside of the stage wall. It is 
really not critical, but it is essentially parallel to the floor. 

QUESTION : Is there not apt to be a tremendous change in level? 

MR. DENNEY: We thought that ought to be true, but surprisingly it is not no- 
ticeable to the people wearing the receivers. I might mention that the receivers 
are worn parallel to the floor, usually in the form of a necklace. 

MR. W. S. STEWART: Why did you choose 100 kilocycles instead of a higher 

MR. DENNEY: Our problem was to concentrate a magnetic field in the interest 
of a minimum amount of radiation. We felt that the longer the wavelength, the 
less radiation there would be in the different sides of the loop. Actually there is 
such a small part of one wavelength in the loop that no standing waves exist. 

QUESTION: For lower power and higher efficiency, would not a somewhat 
higher frequency be better? 

MR. DENNEY : That may be true, but we fixed upon the frequency in our minds. 
We went to work and happily enough it worked very well. On the second unit, 
we hope to make measurements indicating the exact degree of field intensity, and 
so on. 

New Automatic 

Sound Slidefilm System* 



Summary A new completely automatic sound slidefilm system has been 
devised which operates by virtue of low-frequency tones recorded on the disk 
along with the program material. The system permits an uninterrupted flow 
of sound without audible gongs or bells. It is positive in action and immune 
to record wear, scratches, or blunt needles. A tone generator is available to 
record the disks required for the system. 

HAVE BEEN a number of systems proposed from time to 
JL time to synchronize automatically the movement of a film strip 
with a disk recording, thereby eliminating the usual signal bell or 
gong as used on the present sound slidefilm systems. Most of the 
proposed devices have made use of so-called "inaudible" frequencies 
recorded along with the regular program material. A typical device, 
tried some years ago, made use of a 40-cycle tone recorded at the 
proper intervals on the disk to energize a resonant circuit and operate 
a relay, causing the magnetic picture-shifting mechanism to trip. 
This was found to be impractical because it operated purely on a 
marginal basis and if the automatic unit were made sensitive enough 
to be sure to operate from the control tone, it became extremely vul- 
nerable to false tripping by external forces such as accidental shocks, 
flicking of the needle, motor rumble, and other parasitic low-frequency 

Other devices made use of high-frequency tones, usually in the 
6000-cycle region, but this did not prove practical because, since the 
disks are recorded at constant velocity, a 6000-cycle tone must be 
recorded at a very high relative level in order to become at all usable. 
The tone then became almost impossible to eliminate from the audio 
system unless the audio channel was muted for the duration of the 
tone. This defeated the purpose of the automatic device and intro- 
duced an interruption of silence in place of the bell or gong, and re- 
. suited in no great improvement over the manual sound slidefilm. 

* Presented May 18, 1948, at the SMPE Convention in Santa Monica. 



In the development of the 30/50 automatic sound slidefilm system, 
J. T. Mullin has incorporated a new approach which has the advan- 
tages of a tone-on-the-record system and eliminates the erratic 
behavior of the older single tone. In this system, a 50-cycle tone, 
called the "lock-out" tone, is recorded throughout the disk, along with 
the program material, at such a level that it does not cause inter- 
modulation. This 50-cycle tone is used to immobilize or "lock out" 
any tendency for the unit to function except when a picture change is 
desired, then the 50-cycle "lockout" tone is interrupted and a 30- 
cycle or "operate" tone is introduced. This passes readily through 
the automatic circuit, operates a relay, and causes the magnetic 
picture-changing mechanism to function. By having one tone 
(50 cycles) as an inhibiting factor and a second (30 cycles) as a posi- 
tive or "operate" factor, the device becomes completely reliable. 

A projector with the device actually is simpler to operate, than the 
nonautomatic type, since the threading of the projector is the same, 
that is, the strip is placed in the projector and focused on the first 
frame. Then with the turntable running, the needle is slid into the 
first groove of the record. The projector then carries on and shifts 
all the frames automatically at the proper cues. The device will 
not be tripped falsely by flicking the needle nor is it necessary that the 
needle be placed in the very first groove. 

Fig. 1 shows a block diagram of the unit incorporated in a sound 
slide projector. The signal from the pickup, which can be an in- 
expensive crystal cartridge, is fed to a normal volume control from 
which the audio frequencies go to an ordinary amplifier and are de- 
livered to a loudspeaker. It is desirable to provide a low-frequency 
falloff in the audio-frequencj 7 amplifier to minimize any possible 
pickup of the low-frequency tones, although usually the small baffle 
and speaker used will not deliver appreciably anything under 150 
cycles. Another lead from the pickup is connected directly to the 
automatic analyzing circuit which consists first of a low-pass filter 
which eliminates all frequencies above 100 cycles and passes those in 
the 30- to 50-cycle region on to an ordinary amplifying stage. The 
signals then go to another amplifier stage which is driven to saturation 
and becomes a limiting means. Therefore the output from this last 
amplifier will be the same, regardless of which of the two frequencies is 
handled, for the amplification is of such an order that the lowest 




output from the crystal pickup under the most unfavorable tempera- 
ture is sufficient to saturate the amplifier fully. When this limiter 
stage is saturated with the 50-cycle lockout tone, it becomes very in- 
sensitive to any other components of a lower magnitude and any 
30-cycle components from rumble or accidental shocks are compressed 
to the extent that they cannot cause false operation. 

The signal from the limiting amplifier is then carried to a band- 
pass filter which has a high Q in the 30-cycle region and will pass very 
little of the 50-cycle tone. From the band-pass filter the signal 
passes to a rectifier combined* with a time-delay circuit. Thus the 
direct voltage developed after the tuned circuit is dependent only on 

Fig. 1 Block diagram of 30/50 automatic analyzing circuit. 

the frequency of the applied tone as long as the amplitude is sufficient 
to saturate the tube. The direct voltage developed as the result of 
a 50-cycle tone is very low whereas the voltage developed from a 
30-cycle tone is some 30 volts, at least five times as great. This 
differential in voltage, which is independent of the variations in 
record characteristics, pickup temperature variations, and needle 
conditions, is used to operate a relay which in turn causes the opera- 
tion of the magnetic drive on the film-strip projector. 

There have been a number of circuits worked out to accomplish 
this series of functions. In Fig. 2 is shown one satisfactory circuit. 
The signal is fed into the resistance-capacitance network consisting of 
two resistors and two capacitors which results in enough of a rolloff 




beyond approximately 70 cycles, to prevent interference from the 
audio frequencies. The first tube acts as an amplifier and passes the 
signal on to the pentode which receives enough signal to be overdriven 
at all times by either the 30- or the 50-cycle frequencies. The pentode 
then feeds into a transformer which has its secondary tuned by a 
capacitor X to peak at 30 cycles and this acts as the band-pass filter. 
It will pass the 30-cycle components with a great deal more efficiency 
than the 50-cycle ones although the output from the pentode is the 
same regardless of which frequency is being picked up. The capacitor 

Fig. 2 Schematic diagram of 30/50 circuit. 

Y serves as a time-delay means to prevent false operation by an 
occasional transient such as a bump against the machine or the 
flicking of the needle. 

The final tube is biased to cutoff when the direct voltage, developed 
as a result of the 50-cycle lockout tone, is in force and therefore current 
is prevented from flowing in the relay except when the 30-cycle tone 
occurs. At that time, sufficient plate current is developed to close 
the relay contacts positively. 

A quarter-watt neon tube is placed between the screen grid of the 
pentode and ground and serves to maintain the screen at a constant 
potential of 65 volts. In this way the circuit becomes practically 

324 PALMER March 

immune to line-voltage fluctuations, the operation being substantially 
the same from 85 to 140 volts line potential. 

The system requires no change in the technique of the film-strip 
production or printing nor is the pressing of the disks changed in any 
way. The one point requiring the maintenance of certain new 
techniques is in the recording of the disks. 


In order to facilitate the making of records for the 30/50 automatic 
system, a tone generator has been designed to supply the two tones 
and mix them with the program material. This is a photoelectric 
oscillator, making use of a revolving photographic disk, driven by a 
synchronous motor, and having the two tones recorded in concentric 
rings upon it. An electronic switching circuit operates a relay to 
shift the current from one small exciter lamp which generates the 
50-cycle tone, to another similar exciter lamp which generates the 
30-cycle tone. Normally the generator delivers a 50-cycle tone until 
the remote push button is pressed. This shifts the frequency to 30 
cycles for one and one-half seconds. It is then restored to 50 cycles 
automatically. Another push button is located on the panel which 
will cause a 30-cycle tone to be generated as long as the push button 
is pressed. 

In use, the program material is connected to the input of the 
tone generator and the output from the generator is patched to the 
amplifier feeding the disk-cutter recording head. The tone generator 
discards all frequencies below 150 cycles in the audio channel and 
mixes others with the 50- and 30-cycle tones. There are two gain 
controls, one for each of the two tones, since each must be adjusted 
to the proper value, which will depend upon the amplifier and cutter 
characteristics. To simplify the making of this adjustment of rela- 
tive levels between the 30- and 50-cycle tones, a test disk is furnished 
which has the two tones recorded at the optimum level. The two 
gain controls are adjusted to deliver tones at approximately 15 
decibels under full modulation as shown by the volume indicator on 
the" disk cutter. Then test cuts of both frequencies are made and 
compared with the respective levels on the test disk. Any indicated 
level changes are made and further test cuts are compared until they 
are within 1 decibel of the standards on the test record. The re- 
cording can then be made without further problems, the cues being 
placed at the proper intervals in the continuity by depressing the 


remote push button on the end of a portable cord supplied with the 
instrument. It is important that the 50-cycle tone be recorded on the 
disk from the very start as well as continued through to the tailoff 
and closed groove at the finish of the record so that it will serve its 
function as a lockout to prevent false operation. 


By the use of two low-frequency tones, one to serve as an operation- 
preventing factor and the other to insure operation positively, an 
automatic sound slidefilm system has been devised which is com- 
pletely reliable in operation. The elimination of the interrupting 
gong makes possible a vastly improved medium approaching a talking 
motion picture in its ability to sustain audience interest and transfer 
information efficiently. 


Something for Nothing 

The managers of the moving picture shows in the big theaters have 
become so thoroughly imbued with the idea that they should get every- 
thing they want in the way of slides for nothing, that they are the most 
parsimonious lot ever known when it is necessary to buy something. 
Most of them make a cheap show of themselves when they throw an 
announcement on the screen. Instead of buying a beautifully painted 
photographic slide, they use plain glass coated over with opaque, with 
the message scratched through, which to a person who desires to see a 
perfect show causes a thrill of disgust. The managers think because 
some music publisher has furnished them a few sets of song slides free 
they should get announcement slides free also. The meanest looking 
announcement slides, poorly written and almost illegible, are used at 
the Grand Opera House. 

The Moving Picture World, June 18, 1908 

Magnetic Device for Cuing Film 



Summary The magnetic cuing device was designed to eliminate the 
necessity of cutting notches in the edges of motion picture originals used in 
film printers. 

The method of magnetic cuing consists in painting a small dot of magnetic 
material on the edge of the film between two perforation holes. The paint 
may be placed on either the emulsion or base side of the film and still be de- 
tected by the low impedance (50-ohm) magnetic detector. 

The equipment as designed contains two independent cuing channels, one 
for controlling printer-light changes and the other for controlling a fade-in or 
fade-out device which is built in on some printers. A mounting for two ex- 
tremely small pickup heads has been designed to allow direct replacement of 
existing notch actuator switches in film printers. 

THE MAGNETIC CUING device was designed primarily for use on 
16-mm color printers, however, it is believed that it will have use 
in the printing of 35-mm film and possibly also in re-recording of 
sound tracks. 

Present-day requirements of 16-mm printing call for printer-light- 
intensity changes to correct for camera-exposure discrepancies and 
for installation of fade-ins and fade-outs and dissolves in the printer 
rather than in the camera. In Kodachrome-film printing, where it is 
desirable, if not essential, to print each duplicate from the Koda- 
chrome original, two separate sets of cuing marks must be used ; one 
to control the device for light changes and the other to control a 
fade-in and fade-out device. When the problem arose of how to 
cue the fade-in fade-out device, the obvious solution was to use 
the same method of cuing fades as has been used during light 
changes, namely, a notch cut in the edge of the film. However, it 
has been obvious for some time that the method of cutting notches 
even on one edge of the film for controlling light changes had several 
disadvantages. These disadvantages would be more than doubled 
by the necessity of cutting a second set of notches on the other edge of 
the film to control fade-ins and fade-outs. Some of the most ap- 
parent disadvantages of the method of notching films are these : 

* Presented May 18, 1948, at the SMPE Convention in Santa Monica. 



1. Notches once cut in the film are permanent. They cannot be 
removed. If it becomes necessary to make changes in the position 
of the notches, the notched section must be filled in laboriously. 

2. Cutting notches definitely weakens the film and, we believe, 
shortens the working life of the film, thus reducing the maximum 
number of duplicates which can be made from a single original. In 
the commercial- and educational-film field, the need may arise for 
making several hundred copies from a single Kodachrome original. 

Fig. 1 Close-up of two heads mounted on a Bell and Howell Model J 
16-min continuous printer showing relation of pickup heads to film and film- 
guide rollers. 

3. In some printers, the cutting of a notch in the edge of the 
film causes a definite sideways jump of the picture in the frames 
adjacent to the notch. Cutting of notches in both edges of the 
film conceivable could cause sideways jumps in both directions. 

The need for some systems of cuing film, other than cutting notches, 
has long been realized. Two other systems have been developed to 
accomplish the same result^ One uses a separate cue track, running 
at one quarter the speed of the picture original. 

328 LARSEN March 

All the necessary "cuing" notches are cut in the cue track, in syn- 
chronism with the picture. This method requires a special syn- 
chronizer and major modification of the printer to handle the cue 
track. Also, the operation of making up the cue track is slow and 
laborious and is subject to considerably more errors than other 

Another method, which has been described in detail, 1 uses a small 
white dot painted on the film. Then the film is scanned with a light 
beam and a photocell. 

The most recent system, a magnetic cuing device, is believed to be 
simpler than either of these previous systems. In this device, a small 
dot of magnetic material is painted on the edge of the film. Then 

Fig. 2 Amplifier chassis in which all amplifying circuits and the control or 
pilot relay are located. 

the edge of the film is scanned with a magnetic detector or pickup 
head. The resultant electrical pulse is utilized to control electroni- 
cally either light changes or fades or both. 

The physical equipment for magnetic cuing takes the form of 
separate pickup heads operating at 50 ohms output impedance. The 
two pickup heads are shown in Fig. 1, in their relation to the film and 
to the film-guide rollers. One head is used to control light changes 
and the other to control fades. In the photograph, the upper pickup 
head controls light changes and the lower pickup head controls the 
fade-in fade-out device. The pulse-signal output of each pickup 
head is fed to a conventional preamplifier. Signal levels and input- 
transformer shielding conform to present-day high-standard micro- 
phone preamplifier practice. Normally, a 10-foot length of twisted- 
pair shielded cable is utilized for connecting the pickup head to the 


amplifier. The amplifier chassis, in which all amplifying circuits 
and the control or ' 'pilot" relay are located, is shown in Fig. 2. 
However, the amplifier could be located at distances up to 100 feet 
from the pickup head if care is taken in routing of the cable similar to 
that used in routing microphone cable. The input transformer is 
multiple-shielded with permalloy and copper shields to preserve a 
good signal-to-noise level and to allow mounting of two separate pre- 
amplifier channels on the same chassis with the equipment power 

The signal is a low-frequency pulse in the neighborhood of from 
50 to 100 cycles, depending upon printer speed. Since it was found 
that mechanical shock excitation contained considerable high- 
frequency components, the amplifier high-frequency response was 
limited to increase further the operational signal-to-noise ratio and 
to provide a safety factor against mechanical excitation of the pickup 

Following the conventional two-stage, high-gain, signal amplifier, a 
biased clipper tube is provided to reject tube hiss, random noise, 
mechanical shock noises, and to provide a general safety factor and 
sensitivity adjustment. The clipper tube operates in reverse fashion 
to those normally associated with speech circuits. This clipper 
removes and discards the lower portion of the signal and all noise. 
It allows only the signal peak to pass and to actuate the "one-shot" 
multivibrator circuit which operates the control relay. The multi- 
vibrator consists of two triodes in a single envelope. The first triode 
normally conducts continuously and acts as a signal amplifier for 
the small clipped signal. When the multivibrator operation cycle 
begins, the first triode is cut off. The second triode normally is cut 
off and fires only when the multivibrator operates. 

The control relay is in the cathode circuit of the second triode 
and is actuated for an interval of time determined by the constants 
of the multivibrator circuit and not by the amplitude or length of 
the signal pulse. This latter feature removes the random effect of 
varying signal strength and pulse length brought about by different 
size magnetic -paint dots and by placing the pickup heads at varying 
distances from the film. The over-all pickup head and circuit 
sensitivity is such that the pickup head at no time touches the film. 
A spacing of about Vie inch is normally maintained between the 
pickup head and the film. The pickup head is contoured to fit next 
to a roller over which the film passes. The normal film path is not 

330 LARSEN March 

disturbed. In order that chance mechanical motion of the pickup 
head with respect to the near-by mass of the roller does not result in a 
spurious signal, the roller must be made of a nonmetallic material. 
A phenolic linen roller with a metal bearing insert is used to replace 
the usual metal roller. 

On printers having alterriating-current-operated solenoids near the 
film path, where the pickup heads must be installed, it is necessary to 
install Mu-metal or Permalloy slip-on shields over the outer body of 
the solenoids. This is required to overcome interaction of the 
solenoid field with the pickup head. 

The magnetic paint consists of a mixture of hydrogen-reduced 
powdered iron and clear fingernail polish with acetone added. Hy- 
drogen-reduced iron is readily available at any prescription phar- 
macy and is pure iron in its most finely divided commercial state. 
Clear fingernail polish was chosen because it dries very rapidly, is 
easily obtainable, and comes in a small bottle complete with ap- 
plicator brush. When the iron, polish, and acetone are properly 
proportioned, the magnetic paint dries very rapidly, in from 15 to 
20 seconds. This paint adheres to either side of the film, emulsion, 
or base. The paint may be removed from the film by scraping with 
either a razor blade or a retoucher's knife. If the space between the 
cuing dots and the printer aperture can be standardized, it will 
greatly simplify the making of prints when originals are transferred 
from one laboratory to another. This new cuing device offers the 
possibility of setting up standards of spacing which have never existed 
in the notching of films. We suggest that the pickup heads be 
mounted 2 frames apart and that the light change head be set 16 
frames from the printing aperture with the fade-in fade-out head at 
14 frames from the aperture. Also, the light-change dots should be 
placed on the edge of the film opposite to where the sound track 
prints. This would place the fade dots on the same side as the sound 

It is hoped that the introduction of this new magnetic cuing device 
will help to standardize the 16-mm printing work in laboratories all 
over the country. If a standard number of frames between the 
magnetic dot and the printing aperture can be agreed upon, so that a 
film printed in one laboratory can be printed in any other laboratory 
without changing the position of the magnetic dots, a very great 
advance toward standardization and simplification will have occurred. 



(1) Irwin A. Moon, "A photoelectric film cuing system," /. Soc. Mot. Pict. 
Eng,, vol. 49, pp. 364-372; October, 1947. 


MR. LLOYD THOMPSON: Mr. Larsen, what I have is not exactly in the nature 
of a question, but I should like to add some of my own comments to this. We are 
very much in favor of this type of cuing device. As a matter of fact, about a year 
ago, we built up a model like this, or very similar, and tried it out and convinced 
ourselves that it worked very well. There were a few things we did a little differ- 
ently than you did. We tried, for instance, the lacquer with the iron powder, 
which is about the same thing as the fingernail polish. We had some objections 
to it because we found it was possible to peel it off. It would not always peel off, 
but you could not depend that it would always stick, either. We feel that that is 
one of the things that should be standardized before it is used, in some sort of ink 
or paint, that will stay on the film and there will be no question about it. 

Then, we concentrated on a little simpler amplifier than you did. We ended up 
with a one-tube amplifier and we had what we called a reluctance-type circuit. 
We think it would be possible to use some metal, other than steel, so that you 
could buy cellophane that has a coating of iron. It makes a little difference in 
your pickup head, perhaps, and your circuits; or you could use a magnetic device 
as you have suggested, where you actually record a signal on some of this dope 
that they use on magnetic recording of film and have that trip the circuit. Many 
of those things are a little expensive and we prefer to have something that will 
work with a one-tube amplifier if we can, since it would be much cheaper to build 
and simpler to maintain so you can afford to put them on all the printers you have. 

Aside from all of that, we think this question of where the dot should be placed 
on the film should be very definitely standardized before anyone uses it commer- 
cially because anyone with 16-mm laboratory business today knows what a mess 
this notching is. We get films that have been to other laboratories that have 
notches on one side or both sides, and not one set of notches but maybe two or 
three sets of notches. A long time age we decided it would never be possible to 
standardize these notches because. too many people have built their own printers 
and you can never get them all to agree to place them at one place on the film. 
I think that is just out. With this device it seems to us you can place them any- 
where on the film. That is, anybody can take any printer and put this device on 
it, so if you have a standard anyone can use it on any existing printer. 

When we got this experimental model made, instead of putting it in ^ operation 
we thought it should be standardized so we wrote to the SMPE Standards Com- 
mittee, Maurer, Hancock, and Kodak and told them about it and said, "We should 
like for you to get together and set up a standard before we go any further with 
this," and that they agreed to do. As time went on nothing much happened to it, 
but not too long ago John Maurer wrote to me and said, "If you want to get some- 
thing started on standardization you better suggest a standard and let them start 
from there." We did not want to do that. We wanted the Standards Com- 
mittee to originate the standards, but they said somebody had to originate it first. 
So we did a little investigation and wrote to several people that we knew that had 


different types of printers and we came up with a suggested standard of having 32 
frames instead of 16. In that way we figured that it could be used on the present 
Bell and Howell printer, the Depue printers, the printers we use ourselves, and we 
also wrote to George Colburn and he said it would work on his printers. Now, 
that is only a suggestion, but we should like to have the Standards Committee 
take it and set up a standard as to where this mark should be placed, what sort of 
ink or paint should be used so that it is standard; and we have always felt that the 
size of the dot should be standardized. With your device, where you can change 
the amplifier so that it will work at different printer speeds, maybe you have the 
answer to that so that it will only be a dot. We felt maybe it ought to be an 
elongated thing instead of a dot. 

MR. LARSEN: When I said "dot" I was speaking rather loosely. It is actually 
a dash. 

MR. THOMPSON: I know what you mean, because there are different printer 
speeds and there has to be some way of taking care of that because a slow printer 
will actuate it easier than one running fast. 

MR. LARSEN: That is one of the advantages of the more-complicated circuit 
we came up with, because it is entirely independent of the thickness of the dash. 

MR. HUMPHREY: Just how long do these dots stay on? Do you find that they 
have a tendency to wear off? 

MR. LARSEN: No. That is one of the advantages of the concoction we have 
used, the mixture of fingernail polish and acetone. They are very, very perma- 
nent. What actually happens is this : A film is soluble in acetone and when they 
get together they act as they do with film cement. The iron paint actually 
cements itself to the base side if you paint it on the base side, which we do, al- 
though it is not necessary to do so. 

MR. HUMPHREY: I am interested in the positioning of the cuing marks. Evi- 
dently your cuing marks are in the same position as the Bell and Howell notch cuer. 
MR. LARSEN: Yes. As a matter of fact, this device is built so you can remove 
the Bell and Howell notcher and slide this in place. That is the reason we suggest 
using 16 frames. 

MR. HUMPHREY: If you get a film that is already notched you will not have 
much room. 

MR. LARSEN : All we do in that case is to put our paint on the next frame after 
the one that happens to have the notches cut out of it where we want to put them, 
which does not make any difference in the actual result. With this device we can 
take a film that has numerous sets of notches cut in it and ignore all of them and 
put a set of dots on where we have to put it with our particular setup, which 
happens to be 16 frames, and proceed from there, ignoring the previous set of 
notches. "However, we think it is better if the notches are not there in the first 

Improved 35-Mm 
Synchronous Counter* 


Summary The types of synchronous counters, or heads, in general use 
present several problems in the cutting of the multiplicity of negatives used in 
color processes. The necessity of constantly, threading and unthreading 
negatives to match the work print invites scratches and other damage to the 

An improved type of synchronous counter has been developed which 
accommodates either three or six negatives (six being used in dissolves and 
fades) together with a work print. A special positioning lock allows the work 
print to be advanced or retarded in relation to the negative, and relocked in 
frame. To facilitate threading, the keeper rollers are designed to lift into a 
vertical position, thus making it possible simply to set the film onto the sprocket 
without having to slide it under or over rollers or sprockets. A novel ar- 
rangement of a lucite stripping-shoe allows illumination of the film from 
beneath in order to locate frame lines. 

AN IMPROVED 35-mm synchronous counter has been designed and 
put into operation. The design was such as to eliminate some 
of the more objectionable features commonly found in available 
synchronous machines, outstanding among which may be listed : 

1. An arrangement of keeper rollers which makes it necessary to 
slide film under the rollers and over the sprockets when threading or 
unthreading, a situation which is conducive to scratching. 

2. The necessity of unthreading and rethreading the work print 
in order to advance or retard it in relation to the negative being cut. 

3. The lack of proper illumination for viewing the frame lines and 
the subject matter on the film directly over the sprocket. 

Fig. 1 shows the synchronous counter with the keeper rollers 
closed. Fig. 2 is a close-up of the forward end of the machine, with 
the various parts numbered for easy reference. In operation, the 
work print is threaded on the front sprocket (1), after releasing the 
keeper rollers (2). This is done by pressing the spring release (3) for- 
ward. When the film has been placed on the sprocket, the keeper 
rollers are closed simply by pressing them down into position against 

* Presented May 18, 1948, at the SMPE Convention in Santa Monica. 



the film. The rear sprockets (1A) are threaded in the same way. It 
can be seen that with the keeper rollers lifted into a vertical position 
when open, film scratches due to threading are practically eliminated. 
When matching negatives to a cut work print, frame to frame, the 
necessity of removing the work print from the machine at every cut is 
eliminated. First, the work print is rolled down to the desired edge 
number or splice where the negative cut is to be made. Then button 
(8) is pushed, which applies a brake (7) against the inside of sprocket 
(1). This holds the work print in the desired position. Next, handle 
(5) is pulled out, sliding a key out of spline (6), which leaves the 
remainder of the sprockets freewheeling for advancing negatives to 
the place where the cut is to be made. Then handle (5) is pushed in, 

Fig. 1 Four-gang synchronous film counter for cutting multiple-color 


the brake is released, the cut is made, and both work print and 
negatives are rolled to the next scene. The counter, a five-figure 
foot and frame counter commonly used on motion picture equipment 
is geared to the shaft carrying the three rear sprockets (1A). The 
counter may thus be used for counting the footage from the beginning 
of a scene to the point where a negative cut is to be made, and for 
scene-to-scene footage in making continuity cards. 

To obtain a field of illumination directly over the sprockets for 
viewing frame lines and checking subject material, ports were pro- 
vided in the base of the machine so that light from underneath could 
be projected up to the sprockets. To carry the light around the 
sprocket hub and under the film, a strip of lucite (9) was formed into a 




U shape and mounted at each sprocket as shown. The area of the 
lucite strip directly over the sprocket was ground, in order to diffuse 
the light. The lucite member is thus made to serve both as an 
illuminating device, and as a stripper shoe. 

Fig. 3 shows an extension of the design shown in Figs. 1 and 2. 
It is a longer version of the same machine, embodying three additional 
sprockets at the rear. The mechanical make-up of this machine is 

Fig. 2 Close-up of four-gang synchronous film counter showing locking 
hand wheels, -brake lock, keeper rollers, and lucite stripper shoe. 

identical with that of the machine shown in Figs. 1 and 2 with the 
exception that there are six gang sprockets on the rear shaft gear to 
the counter. This facilitates the cutting of dissolves, montages, and 
superimposed titles for foreign release. The counter is used pri- 
marily for taking foot and frame counts and making cue sheets. 

Fig. 4 shows a special machine, designed and built for cutting 
the skip-frame negatives used in color-cartoon photography. It is 
constructed so as to accommodate both two-color and three-color 
successive frame negatives. The work print is threaded on the front 



Fig. 3 Seven-gang synchronous film counter for cutting dissolves, montages, 
and superimposed titles. 

sprocket (IB), and the successive-frame negative on the rear sprocket 
(1). Sprocket (1) is rotated at either two or three times the speed of 
sprocket (IB) by means of a small gear transmission (3). The 
shift from one ratio to the other is accomplished by means of knob 
(4). Counter (7) is connected to sprocket (IB) so that print or screen 
footage is counted. 

A significant interest in the Cinecolor synchronous counters has 
been shown by other studios and laboratories, and several of these 
machines have been built for use in other laboratories. 

Fig. 4 Skip-frame synchronous film counter for cutting either 2-to-l*>r 
3-to-l cartoon negatives. 

Proposed Standards for 16-Mm 
and 8- Mm Picture Apertures 

IN CONNECTION with the broad program of reviewing all standards at 
the close of the war, the Standards Committee of the Society of 
Motion Picture Engineers was asked to restudy the six American 
Standards for picture apertures in 16-mm and 8-mm cameras and 
projectors. Since it appeared that a thorough revision was in order, 
a subcommittee, with John A. Maurer as chairman, was organized for 
this assignment. By the end of 1947, the pattern and most of the 
details of the new proposals were well established. Because of other 
changes at that time, it became desirable to disband the special sub- 
committee and to transfer to the Society's standing Committee on 
16-mm and 8-mm Motion Pictures the task of ironing out the re- 
maining few, but important, controversial points. Agreement was 
reached in October, 1948, and the new proposals were passed along 
to the Standards Committee with a recommendation for favorable 
action. The new proposals, four in number, are shown on the follow- 
ing pages. This is in keeping with the policy of the Standards Com- 
mittee of publishing in the JOURNAL all .proposals involving new ma- 
terial or major revisions before taking action on the question of sub- 
mitting them to the American Standards Association. Your com- 
ments are invited. 

Specifically, the four proposed standards are entitled : 

Z22.7 Location and Size of Picture Aperture of 16-mm Motion Picture 

Z22.8 Location and Size of Picture Aperture of 16-mm Motion Picture 

Z22 . 19 Location and Size of Picture Aperture of 8-mm Motion Picture 

Z22 . 20 Location and Size of Picture Aperture of 8-mm Motion Picture, 


When these are finally approved, they will replace the same Z22 
numbers that were issued in 1941. Since the two 16-mm proposals 
cover sound, as well as silent, equipment, it is intended that they will 
also supplant 1941 American Standards Z22.13 and Z22.14 which 
related to 16-mm sound cameras and projectors. 



In the drafting of these proposed standards, an effort was made to 
dimension the drawings so that they will be of the most direct use to 
the engineer. The introduction of the "K" dimensions, showing the 
distance from the centerline of the aperture to the registering edge of 
the film perforation, is a case in point. 

During the evolution of these proposals, there was a good deal of 
debate on the question of specifying which edge of 16-mm film should 
be guided. The Committee on 16-mm and 8-mm motion pictures 
finally concluded that this question properly should be left to the de- 
signer, but the proposal does call attention to the factors involved. 
A similar statement is made relative to the problem of assigning a 
definite value to dimension "G." . 

In the past, standards dealing with this type of subject matter 
generally have been limited to dimensioned drawings. The Stand- 
ards Committee feels it is desirable to include explanatory text or 
notes to make the intention and application of the standard more 
certain. This practice is followed in these four proposals to a degree 
that makes any further discussion of technical points appear to be 

SEVERAL STANDARDS developed by subcommittees of ASA Sec- 
tional Committee Z10 listed below are available. Authors are 
encouraged to follow these standard symbols and abbreviations. 

Title of Standard Price 

Abbreviations for Scientific and Engineering Terms Z10. 1-1941 $0.45 

Letter Symbols for Hydraulics Z10. 2-1942 0.45 

Letter Symbols for Mechanics of Solid Bodies Z10 . 3-1948 . 30 

Letter Symbols for Heat and Thermodynamics Z10. 4-1943 0.65 

Letter Symbols for Physics Z10 . 6-1948 1 . 00 

Letter Symbols for Chemical Engineering Z10 . 12-1946 . 50 




Proposed American Standard 
Location and Size of Picture Aperture of 

16-Millimeter Motion Picture Cameras 


Page 1 of 3 pages 

This standard applies to both silent and sound 16-mm. motion picture 
cameras. It covers the size and shape of the picture aperture and the rela- 
tive positions of the aperture, the optical axis, the edge guide, and the film 
registration device. The notes are a part of this standard. 


ft- -fS NOTE 1) 







-*- -A 


-C > 

1 \ 













A (measured perpen- 

dicular to edge of 


0.201 minimum 

5.1 1 minimum 


B (measured parallel 

n 000 + 0.006 

7 , 9 +0.18 

to edge of film) 

"' zyz ~ 0.002 

' ' 4 -0.05 



0.31 4 0.002 % 

7.98 0.05 



0.1 10 minimum 

2.79 minimum 



0.1 25 0.002 

3.1 8 0.05 



0.1 75 0.002 

4.44 0.05 



0.474 0.002 

12.04 0.05 



0.773 0.002 

19.63 0.05 



1.072 0.001 

27.23 0.03 



0.020 maximum 

0.51 maximum 



Proposed American Standard 
Location and Size of Picture Aperture of 

16-Millimeter Motion Picture Cameras 


February 1949 

Page 2 of 3 pages 

The angle between the vertical edges of the aperture and the edges of 
normally positioned film shall be degrees, Va degree. 

The angle between the horizontal edges of the aperture and the edges of 
normally positioned film shall be 90 degrees, } /2 degree. 

Note 1 : Dimensions A, B, and R apply to the size of the image at the plane 
of the emulsion; the actual picture aperture has to be slightly smaller. The 
exact amount of this difference depends on the lens used and on the sep- 
aration (dimension G) of the emulsion and the physical aperture. G should 
be no larger than is necessary to preclude scratching of the film. The 
greatest difference between the image size and aperture size occurs with 
short focal-length, large diameter lenses. 

Dimensions A and B are consistent with the size of the images on a 
1 6-mm. reduction print made from a 35-mm. negative with the standard 
2.15 reduction ratio. 

It is desirable to hold the vertical height of the actual aperture to a 
value that will insure a real (unexposed) frameline. This results in less 
distraction when the frameline is projected on the screen than is the case 
when adjacent frames overlap. 

Note 2: The edge guide is shown on the sound-track edge. This location for 
it has the advantage that the rails bearing on the face of the film along 
this edge and also between the sound track and picture area can be of 
adequate width. Disadvantages of this location for the edge guide are 
that, because film shrinkage and tolerances affect the lateral position of 
the perforations, the pulldown tooth must be comparatively narrow and 
will not always be centered in the perforation. 

The guide can be on the other edge, adjacent to the perforated edge 
of sound film. With the guide at this edge, the width of the pulldown tooth 
does not have to be decreased to allow for shrinkage. However, because 

- of variqtions introduced by shrinkage of film, this location for the edge 
guide has the important disadvantage that it makes extremely difficult 
the provision of rails of adequate width to support the sound-track edge 
without encroaching on, and consequently scratching, the picture or 
sound-track area. (See Section 3, Proposals for 16-mm. and 8-mm. 
Sprocket Standards, Vol. 48, No. 6, June 1947, Journal of the Society of 
Motion Picture Engineers). 


Proposed American Standard 
Location and Size of Picture Aperture of 

16-Millimeter Motion Picture Cameras 


Page 3 of 3 pages 

The film may be pressed against the fixed edge guide by a spring, by 
the tendency of the film to tilt in the gate, or by other means. In the second 
case, there is a fixed guide for each edge of the film. The important point 
is to have the film centered laterally on the optical axis. 

Dimension C is made slightly less than half the width of unshrunk film 
so that the film will be laterally centered if it has a slight shrinkage at the 
time it is run in the camera. This is the normal condition. As indicated by 
the above discussion, C may be measured in either direction from the 
vertical centerline. 

Note 3: Dimension F must be maintained only when a photographic sound 
record is to be made on the film that passes through the camera; other- 
wise F may be disregarded. 

Note 4: The K dimensions are measured along the path of the film from the 
horizontal centerline of the aperture to the stopping position of the regis- 
tration device. Both the dimensions and tolerances were computed to 
keep the frameline within 0.002 to 0.005 inch of the centered position for 
films having shrinkages of 0.0 to 0.5 per cent at the time they are exposed 
in the camera. For any given camera, use the value of K corresponding to 
the location of the registration device. 

If the film does not stop exactly where the film registration device leaves 
it, because of coasting or some other cause, a slight adjustment of the 
value of K will be necessary. This will be indicated if film that has a shrink- 
age of 0.2 to 0.3 per cent when it is run in the camera does not show a 
properly centered frameline. From such a test, the amount and direction 
of the adjustment can be determined. 

Note 5: "Optical axis of camera" is defined as the mechanical axis or cen- 
terline of the sleeve or other device for holding the picture-taking lens. 
Except for manufacturing tolerances, it coincides with the optical axis 
of the lens. 




Proposed American Standard 
Location and Size of Picture Aperture of 

16-Millimeter Motion Picture Projectors 


February 1949 

Page 1 of 3 pages 

This standard applies to both silent and sound 16-mm. motion picture 
projectors. It covers the size and shape of the picture aperture and the rela- 
tive positions of the aperture, the optical axis, the edge guide, and the film 
registration device. The notes are a part of this standard. 










A (measured perpen- 

dicular to edge of 


0.380 0.002 

9.65 0.05 


B (measured parallel 

to edge of film) 

0.284 0.002 

7.21 0.05 



0.31 4 0.002 

7.98 0.05 



0.1 24 0.005 

3.15 0.13 



0.1 74 0.005 

4.42 0.13 



0.473 0.005 

12.01 0.13 


K 3 

0.771 0.005 

19.58 0.13 



1.070 0.005 

27.1 8 0.1 3 



1.368 0.005 

34.75 0.1 3 



0.020 maximum 

0.51 maximum 



Proposed American Standard 
Location and Size of Picture Aperture of 

16-Millimeter Motion Picture Projectors 

February 1 949 

Page 2 of 3 pages 

The angle between the vertical edges of the aperture and the edges of 
normally positioned film shall be degrees, Vi degree. 

The angle between the horizontal edges of the aperture and the edges of 
normally positioned film shall be 90 degrees, !/2 degree. 

Note 1 : Dimensions A, B, and R apply to the portion of the image on the film 
that is to be projected; the actual opening in the aperture plate has to be 
slightly smaller. The exact amount of this difference depends on the lens 
used and on the separation (dimension G) of the emulsion and the physical 
aperture. To minimize the difference in size and make the image of the 
aperture as sharp as practicable on the screen, G should be no larger 
than is necessary to preclude scratching of the film. When the reduction 
in size from the image to the actual aperture is being computed, it is sug- 
gested a 2-inch f/1 .6 lens be assumed unless there is reason for doing 

Note 2: The limiting aperture is shown as being between the film and the 
light source so that it will give the maximum protection from heat. If other 
factors are more important, it may be on the other side of the film. 

Note 3: The edge guide is shown on the sound-track edge. This location for 
it has the advantage that the rails bearing on the face of the film along 
this edge and also between the sound track and picture area can be of 
adequate width. Disadvantages of this location for the edge guide are 
that, because film shrinkage and tolerances affect the lateral position of 
the perforations, the pulldown tooth must be comparatively narrow and 
will not always be centered in the perforation. Also, in some prints the 
sound-track edge is slit after processing, in which case there is likely to 
be some lateral weave between this edge and the pictures. 

The guide can be on the other edge, adjacent to the perforated edge 
of sound film. With the guide at this edge, the width of the pulldown tooth 
does not have to be decreased to allow for shrinkage. Also, slitting the 
sound-track edge after processing will not introduce lateral unsteadiness. 
However, because of variations introduced by shrinkage of film, this 
location for the edge guide has the important disadvantage that it makes 
extremely difficult the provision of rails of adequate width to support the 


Proposed American Standard 
Location and Size of Picture Aperture of 

16-Millimeter Motion Picture Projectors 

February 1949 

Page 3 of 3 pages 

sound-track edge without encroaching on, and consequently scratching, 
the picture or sound-track area. (See Section 3, Proposals for 1 6-mm. and 
8-mm. Sproket Standards, Vol. 48, No. 6, June 1947, Journal of the 
Society of Motion Picture Engineers). 

The film may be pressed against the fixed edge guide by a spring, 
by the tendency of the film to tilt in the gate, or by other means. In the 
second case, there is a fixed guide for each edge of the film. The important 
point is to have the film centered laterally on the optical axis. 

Dimension C is made slightly less than half the width of unshrunk film 
so that the film will be laterally centered if it has a slight shrinkage at 
the time it is run in the projector. This is the normal condition. As indicated 
by the above discussion, C may be measured in either direction from the 
vertical centerline. 

Note 4: The K dimensions are measured along the path of the film from the 
horizontal centerline of the aperture to the stopping position of the regis- 
tration device. It is customary to provide a framing movement of 0.025 
inch above and below this nominal position. For any given projector, use 
the value of K corresponding to the location of the registration device. 

If the film does not stop exactly where the film registration device leaves 
it, because of coasting or some other cause, a slight adjustment of the 
value of K will be necessary. 

Note 5: "Optical axis of projector" is defined as the mechanical axis or 
centerline of the sleeve for holding the projection lens. Except for manu- 
facturing tolerances it coincides with the lens axis. 




Proposed American Standard 
Location and Size of Picture Aperture of 

8-Millimeter Motion Picture Cameras 

February 1 949 

Page 1 of 2 Pages 

This standard applies to 8-mm. motion picture cameras. It covers the 
size and shape of the picture aperture and the relative positions of the aper- 
ture, the optical axis, the edge guide, and the film registration device. The 
notes are a part of this standard. 



FILM- (see NOTE ^) 







Ai (measured perpen- 

dicular to edge of 


0.094 min., 0.1 04 max. 

2.39 min., 2.64 max. 



0.094 min. 

2.39 min. 


B (measured parallel 

ni o ft + 0.008 

351 +0 - 20 

to edge of film) 


3 ' 51 -0.03 



0.205 0.002 

5.21 0.05 



0.050 0.002 

1.27 0.05 



0.1 00 0.002 

2.54 0.05 


K 2 

0.249 0.002 

6.32 0.05 



0.399 0.002 

10.1 3 0.05 



0.549 0.002 

13.94 0.05 



0.698 0.002 

17.73 0.05 



0.848 0.002 

21. 54 0.05 



0.998 0.002 

25.35 0.05 



0.010 maximum 

0.25 maximum 



Proposed American Standard 
Location and Size of Picture Aperture of 

8-Millimeter Motion Picture Cameras 

Z 2 2.19- 
February 1949 

Page 2 of 2 Pages 

The angle between the vertical edges of the aperture and the edges of 

normally positioned film shall be degrees, ~- 1/2 degree. 

The angles between the horizontal edges of the aperture and the edges 

of normally positioned film shall be 90 degrees, Vi degree. 

Note 1 : Dimensions A, B, and R apply to the size of the image at the plane of 
the emulsion; the actual picture aperture has to be slightly smaller. The 
exact amount of this difference depends on the lens used and on the 
separation (dimension G) of the emulsion and the physical aperture. G 
should be no larger than is necessary to preclude scratching of the film. 
The greatest difference between the image size and aperture size occurs 
with short focal-length, large diameter lenses. 

It is desirable to hold the vertical height of the actual aperture to a 
value that will insure a real (unexposed) frameline. This results in less 
distraction when the frameline is projected on the screen than is the case 
when adjacent frames overlap. 

Note 2: The film may be pressed against the fixed edge guide by a spring, 
by the tendency of the film to tilt in the gate, or by other means. In the 
second case (generally used in pre-loaded magazines), there is a fixed 
guide for each edge of the film. The important point is to have the film 
located in the correct lateral position with respect to the optical axis. 
The value of dimension C has been chosen on the assumption that the 
film will have a slight shrinkage when it is run through the camera. This 
is the normal condition. 

Note 3: The K dimensions are measured along the path of the film from the 
horizontal centerline of the aperture to the effective stopping position of 
the registration device. Both the dimensions and tolerances were com- 
puted to keep the frameline within 0.002 to 0.005 inch of the centered 
position for films having shrinkages between 0.0 and 0.5 per cent at the 
time they are exposed in the camera. For any given camera, use the value 
of K corresponding to the location of the registering device. 

If the film does not stop exactly where the film registration device leaves 
it, because of coasting or some other cause, a slight adjustment of the 
value of K will be necessary. This will be indicated if film that has a shrink- 
age of 0.2 to 0.3 per cent when it is run in the camera does not show a 
properly centered frameline. From such a test, the amount and direction 
of the adjustment can be determined. 

Note 4: "Optical axis of camera" is defined as the mechanical axis or cen- 
terline of the sleeve or other device for holding the picture-taking lens. 
Except for manufacturing tolerances, it coincides with the optical axis 
of the lens. 




Proposed American Standard 
Location and Size of Picture Aperture of 

8-Miilimeter Motion Picture Projectors 

February 1 949 

Page 1 of 2 Pages 

This standard applies to 8-mm. motion picture projectors. It covers the 
size and shape of the picture aperture and the relative positions of the aper- 
ture, the optical axis, the edge guide, and the film registration device. The 
notes are a part of this standard. 












A (measured perpen- 

dicular to edge of 


0.1 72 0.001 

4.37 0.03 


B (measured parallel 

to edge of film) 

0.1 29 0.001 

3.28 0.03 



0.205 0.002 

5.21 0.05 



0.050 0.005 

1.27 0.1 3 



0.1 00 0.005 

2.54 =t 0.13 


K 2 

0.249 0.005 

6.32 0.13 



0.398 0.005 

10.11 0.13 



0.547 0.005 

13.89 0.13 



0.696 0.005 

17.68 0.1 3 


K 6 

0.846 0.005 

21.49 0.13 



0.995 0.005 

25.27 0.1 3 



1.1 44 0.005 

29.06 0.1 3 



0.010 maximum 

0.25 maximum 



Proposed American Standard 
Location and Size of Picture Aperture of 

8-Millimeter Motion Picture Projectors 

February 1949 

Page 2 of 2 Pages 

The angle between the vertical edges of the aperture and the edges of 
normally positioned film shall be degrees, '/ 2 degree. 

The angle between the horizontal edges of the aperture and the edges 
of normally positioned film shall be 90 degrees, =t '/2 degree. 

Note 1 : Dimensions A, B, and R apply to the portion of the image on the film 
that is to be projected; the actual opening in the aperture plate has to 
be slightly smaller. The exact amount of this difference depends on the 
lens used and on the separation (dimension G) of the emulsion and the 
physical aperture. To minimize the difference in size and make the image 
of the aperture as sharp as practicable on the screen, G should be no 
larger than is necessary to preclude scratching of the film. When the 
reduction in size from the image to the actual aperture is being computed, 
it is suggested a 1 -inch f/1 .6 lens be assumed unless there is reason for 
doing otherwise. 

Note 2: The limiting aperture is shown as being between the film and the 
light source so that it will give the maximum protection from heat. If other 
factors are more important, it may be on the other side of the film. 

Note 3: In 8-mm. projectors the edge guide should bear on the edge of the 
film adjacent to the perforations. The other edge of the film usually is slit 
after processing and so is more likely to weave laterally with respect to 
the pictures. 

The value of dimension C has been chosen so that film having a slight 
shrinkage when it is projected will be properly centered. This is the normal 

Note 4: The K dimensions are measured along the path of the film from the 
horizontal centerline of the aperture to the stopping position of the regis- 
tration device. It is customary to provide a framing movement of approxi- 
mately 0.025 inch above and below this nominal position. For any given 
projector, use the value of K corresponding to the location of the regis- 
tration device. 

If the film does not stop exactly where the film registration device leaves 
it, because of coasting or some other cause, a slight adjustment of the 
value of K will be necessary. 

Note 5: "Optical axis of projector" is defined as the mechanical axis or 
centerline of the sleeve for holding the projection lens. Except for manu- 
facturing tolerances, it coincides with the lens axis. 

65th Semiannual Convention 

Hotel Statler April 4-8, 1949 New York, N. Y. 


EARL I. SPONABLE '..... President 

LOREN L. RYDER Past-President 

PETER MOLE Executive Vice-President 

JOHN A. MAURER Engineering Vice-President 

CLYDE R. KEITH Editorial V ice-President 

DAVID B. JOY Financial V ice-President 

WILLIAM C. KUNZMANN Convention Vice-President 



General Office, New York 

BOYCE NEMEC Executive Secretary 

HELEN M. STOTE Journal Editor 

WILLIAM H. DEACY, JR Staff Engineer 

SIGMUND M. MUSKAT Office Manager 

Chairmen of Committees for the Convention Program 

Convention Vice-President WILLIAM C. KUNZMANN 

Atlantic Coast Section and Local Arrangements .... WILLIAM H. RIVERS 
Papers Committee Chairman NORWOOD L. SIMMONS, JR. 


Publicity Committee HAROLD DESFOR 


Registration and Information WILLIAM C. KUNZMANN 

Assisted by ERWIN R. GEIB, 

Luncheon and Banquet OSCAR F. NEU 

Vice-Chairman LESTER B. ISAAC 

Public- Address Equipment R. E. WARN 

Hotel and Transportation WILLIAM F. JORDAN 

Ladies' Reception Committee . . . - MRS. EARL I. SPONABLE, 



Membership and Subscription . LEE JONES 

Projection Program 35-Mm HENRY F. HEIDEGGER 

* Assisted by officers and members New York Local 306, 1.A.T.S.E. 
Projection Program 16-Mm FRANK B. ROGERS, JR. 




Hotel Reservations and Rates 

The Hotel Statler (formerly Hotel Pennsylvania) will be the headquarters of 
the 65th Semiannual Convention. Room reservation cards were mailed to the 
membership early in February. If you plan to attend the Convention, check 
your desired accommodations and return the card immediately to the hotel so your 
reservation can be booked and confirmed by the hotel management. 

Booked reservations are subject to arrival date change or cancellation prior 
to March 20. 

Rail, Pullman, and Plane Travel 

Eastern travel conditions still remain acute; therefore if attending the Con- 
vention in New York, book your travel accomodations at least a month prior to 
April 4, with your local travel agent. 

Convention Registration and Papers Program 

Members and others within the motion picture industry who are contemplating 
presenting papers during the 65th Semiannual Convention dates are urged to sub- 
mit the title of the paper to be presented, name of author, and a complete manu- 
script to the JOURNAL Editor, Society of Motion Picture Engineers, 342 Madison 
Avenue, New York 17, N. Y. A second copy, without illustrations, should be 
sent to Mr. E. S. Seeley, Altec Service Corporation, 161 Sixth Avenue, New York 
13, N. Y. 

The Convention Registration Headquarters will be located in Conference Room 
9, on the 18th floor of the hotel; all business and technical sessions will be held 
in the Salle Moderne, on the same floor. 

The Publicity Committee and Press Headquarters will also be located on this 

You can register at the Registration Headquarters on the morning of April 4, 
also procure your desired luncheon tickets prior to 10:30 A.M. on this date, to be 
assured table seating. 

Your registration is requested since this revenue from collected registration 
fees is used to defray the Convention expenses. 

Convention Get-Together Luncheon 

The usual convention get-together luncheon will be held in the Georgian Room 
of the hotel on Monday, April 4, at 12:30 P.M. Eminent speakers will address the 
luncheon gathering. 

Most Important Procure your desired luncheon tickets at the Registration 
Headquarters prior to 10:30 A.M. on April 4, otherwise no table seating is guar- 

Checks or money orders for registration fees and luncheon or banquet tickets 
should be made payable to W. C. Kunzmann, Convention Vice-President, and 
not to the Society. 


65th Semiannual (Informal) Banquet and Cocktail Hour 

The Convention's social cocktail hour for holders of banquet tickets will be held 
in the hotel's Georgian Room foyer, on Wednesday evening, April 6, between 
7:15 and 8: 15 P.M. 

The 65th Semiannual Convention informal banquet (dress optional) will be 
held in the Georgian Room on Wednesday evening, April 6, promptly at 8:30 
P.M. There will be dancing and entertainment. Tables for the banquet can be 
reserved at the registration headquarters. 

Ladies' Reception and Registration Headquarters 

The Ladies' Reception and Registration Headquarters will be located in Room 
129 in the hotel and open daily during the Convention. Mrs. Earl I. Sponable 
will serve the Convention as hostess and Mrs. William H. Rivers, cohostess, to 
the ladies attending the Convention. 

The ladies' entertainment program will be announced in later released conven- 
tion bulletins. 

Motion Pictures and Recreation 

The Convention-issued identification cards to registered members and guests 
will be honored at the following de luxe motion picture theaters in New York 
during the Convention : Capitol, Paramount, Radio City Music Hall, Roxy, and 
Warner Strand. 

Literature and information pertaining to places of interest to visit in New York 
and vicinity can be obtained at the hotel's information bureau or the Convention 
registration headquarters. 


Monday, April 4, 1949 

9 : 30 A.M. Registration, Conference Room 9, 18th floor 
Advance sale Luncheon and Banquet tickets 
Procure your luncheon tickets prior to 10:30 A.M., this date 
12:30 P.M. Get-Together Luncheon, Georgian Room 
3:00 P.M. Films in Television (Forum), Salle Moderne 
8:00 P.M. Theater Television Demonstration, Georgian Room 

Tuesday, April 5, 1949 

9:30 A.M. Registration, Conference Room 9, 18th floor 

Advance sale of Banquet tickets 
10 : 00 A.M. Television Session, Salle Moderne 
2:00 P.M. Television Session, Salle Moderne 



Wednesday, April 6, 1949 

9: 30 A.M. Registration, Conference Room 9, 18th floor 

Advance sale of Banquet tickets 

10:00 A.M. High-Speed Photography Session, Salle Moderne 
2:00 P.M. High-Speed Photography Session and Equipment Demonstration, 

Salle Moderne 

7:15 P.M. Cocktail Hour, Georgian Room foyer 
8:30 P.M. Informal Banquet, Georgian Room (dress optional). Dancing and 


Thursday, April 7, 1949 

2:00 P.M. Films and Film Laboratories, Discussion of 16-Mm Reproducer 

Characteristics; Sound Recording, Salle Moderne 
8:00 P.M. Popular Lecture and Demonstration, Salle Moderne 

Friday, April 8, 1949 

9:30 A.M. Registration, Conference Room 9, 18th floor 

10:00 A.M. Theaters and Projection, Salle Moderne 

2:00 P.M. Photographic Equipment and Studio Techniques, Salle Moderne 

5:00 P.M. Adjournment 


Most printing costs in recent years have advanced markedly and 
Society publications are not immune because increased labor charges 
and the cost of paper to our printer have been passed on and are 
charged against the JOURNAL. The Society works on a remarkably 
tight publication budget and it has been necessary to increase the 
JOURNAL subscription rate to nonmember subscribers from $10.00 
to $12.50 per year. All subscriptions renewed prior to March 15, 
1949, were billed at the old rate but subscriptions received after 
March 15 are being charged for at the $12.50 rate. In 1949 we 
expect to print 1622 JOURNAL pages as against 1376 pages for 1948, 
1264 pages for 1947, and 1200 pages for 1946. 

A variety of cost-cutting methods are being investigated with the 
hope that further advances in printing costs may be offset and make 
it unnecessary for additional increases in our subscription rates. 


CHARLES G. WEBER, assistant chief of the Paper Section, 
National Bureau of Standards, died on January 18, 1949. 

He was graduated from the New York State College of Forestry 
of Syracuse University in 1916, and was a veteran of World War I. 

A member of the Bureau's staff for 20 years, Mr. Weber directed 
researches in papermaking materials and processes, printing, pack- 
aging materials, low-cost housing materials, the standardization of 
testing methods for paper and related products, and the use of mo- 
tion picture films for records. 

During World War II, he assisted the Army Map Service in the 
development of wet-strength map paper for field use; he was suc- 
cessful in the development of improved methods in multicolor offset 
printing that greatly reduces losses caused by misregister of the 
prints. He supervised research on low-cost housing materials in 
co-operation with the Federal Housing Administration, and at the 
time of his death, he was directing research on the melamine-resin 
bonding of offset papers. 

Mr. Weber served on standardizing committees of the Technical 
Association of the Pulp and Paper Industry, the American Society 
for Testing Materials, and the Society of Motion Picture Engineers. 
He was a member of the Technical Committee of the Lithographic 
Technical Foundation, and an Active Member of the SMPE. 


Section Meeting 


R. T. Van Niman presided over the December 9, 1948, meeting of the Central 
Section, which was held at the Lincolnwood Plant of the Bell and Howell Com- 
pany. Nearly 200 members, guests, and members of the new optical group were 
present. Charles E. Phillimore welcomed the group and invited those interested 
to a tour of the plant following the meeting. 

The first paper on "Design Considerations for Television Studio Motion Picture 
Projectors," by James V. Starbuck and Elmer Enke, projection engineers, WGN- 
TV, and H. J. Daly and Dudley A. Howell, projectionists, television station 
WBKB, was read by Mr. Howell. He outlined the arrangement and problems of 
the television projection rooms of WBKB and WGN-TV. He indicated that in 
the main, the 35-mm equipment was satisfactory but with the syncrolite lamp 
source it was impossible to frame the picture during projection. The 16-mm 
projection equipment has many faults to make it a practical instrument in the 
projection booth. These faults were outlined as follows: (a) equipment too 
light, (b) needs real focusing adjustment, (c) sprockets too small, (d) should have 
brake for instantaneous stopping, (e) replacement of lamp should be easier, a 
matter of seconds instead of minutes, (f ) douser should be developed, (g) improved 
mounting for preamplifier is needed, and (h) claw intermittent should be 

"Continuous Stereoscopic Aerial Strip Camera Photography," was presented 
by Colonel George W. Goddard, chief of the Photographic Laboratory, United 
States Air Force, Wright Field, Dayton, Ohio. With a series of excellent slides, 
Colonel Goddard traced the history of Aerial Reconnaissance Photography from 
1880 to the present time. This included the planes used up to the very latest 
designs of jet planes flying at 420 miles per hour photographing at an altitude of 
4000 feet; 240-inch lenses are used. In the model now being developed the 
camera will use 40-inch wide film in 400-foot lengths. 

The last part of Colonel Goddard's talk was a demonstration of stereoscopic 
color photography taken from the air with a slit camera. The film travel in this 
camera is automatically governed by the ground speed of the plane, insuring high 
resolution of the twin images. Some very startling and interesting scenes of the 
cities of Germany were viewed with the use of polaroid glasses. These films were 
projected from the rear on a 12- by 16-foot special screen. 

Armed Forces 
Communications Association 

Brigadier General David Sarnoff, president of the Armed Forces Communica- 
tions Association and also chairman of the Board of the Radio Corporation of 
America, announced that the third annual meeting of the Armed Forces Com- 
munications Association will be held in Washington, March 28 and 29, 1949. 

Business meetings will be held the first day and will be climaxed by the annual 
banquet at which it is expected nearly 1000 members will attend. This year's 
meeting will feature the Navy's communications and photographic activities. 
Navy leaders and other distinguished government figures will be the principal 
speakers at the banquet. The second day and perhaps part of a third will be 
devoted entirely to exhibits and demonstrations planned and directed by the 
Navy at its stations and aboard ships in the Washington area. 

The Association, made up of civilians and military members, is dedicated to 
the purpose of insuring that the Navy, Army, and Air Force will have the best 
in communications, radar, and photography. 


Monday, March 28 

9:00 A.M. 
REGISTRATION Shoreham Hotel. 

10:00 A.M. 

ham Hotel. 

12: 00-1: 30 P.M. 
LUNCHEON Shoreham Hotel. 

1:30-4: 00 P.M. 

members. Addresses by the Chief 
of Naval Communications, the Chief 
Signal Officer, the Director of Air 
Communications, the President of 
AFCA, and his successor. Presenta- 
tion of AFCA certificates of merit. 
Orientation by Navy of exhibits and 
demonstrations . 

6:00 P.M. 
COCKTAILS Shoreham Hotel. 

7: 30 P.M. 

BANQUET Shoreham Hotel. Ad- 
dresses by Admiral Louis E. Denfeld, 
Chief of Naval Operations, and by 
AFCA's president, David Sarnoff. 

Tuesday, March 29 

ARRANGEMENTS for the second day's 
meeting will all be made by the Navy 
Department, Captain Robert J. Foley, 
USN, in charge. 

9: 30 A.M. 

TRANSPORTATION from Shoreham Hotel 
to Navy Exhibits. 

10: 00 A.M. 

EXHIBITS of Naval communications 

1:00 P.M. 
LUNCHEON Navy Station. 

2: 30-5: 00 P.M. 

DEMONSTRATION of Navy communi- 
cation, photographic, and combat 


1949 Nominations 

THE 1939 NOMINATING COMMITTEE, as appointed by the President of the 
Society, was confirmed by the Board of Governors at its January meeting. 

D. E. HYNDMAN, Chairman 

Room 626 342 Madison Avenue 

New York 17, N. Y. 


General Precision Equipment Corp. Technicolor Motion Picture Corp. 

63 Bedford Road 6311 Romaine Street 

Pleasantville, N. Y. Los Angeles 28, Calif. 

Research Laboratories L^T^^ 

National Carbon Company ^97 Fifth Avenue 

Box 6087 New York 17, N.Y. 

Cleveland 1, Ohio T T GoLDSMITH 

F. E. CAHILL, JR. Allen B. Du Mont Laboratories 

Warner Bros. Pictures, Inc. 2 Main Avenue 

321 W. 44 Street Passaic,N.J. 
New York 20, N. Y. 


General Electric Company Western Electric Company 

Nela Park 6601 Romaine Street 

Cleveland 12, Ohio Hollywood 38, Calif. 

All voting members of the Society who wish to submit recommendations for 
candidates to be considered by the Committee as possible nominees, are requested 
to correspond directly with the Chairman or any of the members of the Nominat- 
ing Committee. Active, Fellow, or Honorary Members are authorized to make 
these suggestions which must be in the hands of the Committee by May 1, 1949. 

There will be eight vacancies on the Board of Governors as of January 1, 1950, 
which must be filled. Those members whose terms of office expire are: 

Financial Vice-President D. B. JOY 

Engineering Vice-President J. A. MAURER 

Treasurer R. B. AUSTRIAN 

Governor. . . .A. W. COOK Governor. . . .P. J. LARSEN 

Governor. . . . JAMES FRANK, JR. Governor. . . .G. E. SAWYER 

Governor . . . . L. T. GOLDSMITH 

The recommendations of the Nominating Committee will be submitted to the 
Board of Governors for approval at the July meeting. The ballots will then be 
prepared and mailed to the voting members of the Society forty days prior to the 
Annual Meeting of the Society. This is the business session held during the Fall 
Convention, which this year will be in Hollywood, California, October 10-14. 

D. E. HYNDMAN, Chairman, 
Nominating Committee 

Book Reviews 

Sound and Documentary Film, by K. Cameron (Foreword by 

Published (1947) by Sir I. Pitman and Sons, Ltd., Pitman House, 39-41 Parker 
Street, Kingsway, London, W.C. 2, England. Also distributed by Pitman 
Publishing Corporation, 2 W. 45 St., New York, N. Y. 157 pages + XV pages + 
3-page index. 77 illustrations and diagrams. 5 l / 2 X 8 l /z inches. Price, 15 

This little book of 157 pages represents an analysis of some of the problems that 
face the producer and the sound engineer when making a documentary film. 
While the primary emphasis is on British films, the discussion can be applied 
freely to films made in the United States. According to the author, "in the per- 
fect sound film the actual sound should be so perfectly wedded with the picture 
that the illusion of reality is complete." Documentary films, as we know them 
today, are about twenty years old. Two general classes of documentary films are 
described: (1) the straightforward description of an incident with a simple com- 
mentary, music, and sound effects, and (2) an imaginative, human exposition of 
how ordinary people live and work. The latter type is more difficult to make 
well, and is represented by such films as "Target for Tonight" and "Listen to 
Britain." The book describes the planning of a documentary or "realist" film, 
the problems facing the sound crew, the use of music and sound effects, post- 
synchronizing and dubbing, re-recording, and finally showing the film. The last 
50 pages are devoted to brief technical abstracts of some of the processes involved 
in recording sound, and to a glossary of technical terms. 


Eastman Kodak Company 

Rochester, N. Y. 

Discharge Lamps, by H. K. Bourne 

Published (1948) by Chapman and Hall, Ltd., 37 Essex St., W.C. 2, London , 
England. 417 pages + 7 pages -f XV pages. 186 figures. 5 3 /4 X 8 3 /4 inches. 
Price, $12.00. Book available from American Photographic Publishing Co., 
353 Newbury St., Boston 15, Mass. 

The reader who is primarily interested in the characteristics of light sources 
will find this book a convenient reference. As the title indicates, the author is 
principally concerned with discharge sources, but he also describes in considerable 
detail the characteristics of tungsten-filament lamps, carbon arcs, and photoflash 
lamps. One may assume that data on these types are included in order better to 
establish the effectiveness of discharge sources for many photographic and cer- 
tain projection applications. 

While discharge lamps may be constructed employing any of several gases, it is 
appropriate that lamps employing mercury vapor should be the principal theme 
of this book. The relatively high efficiency of such sources, particularly in terms 


Book Reviews 

of their effect on photographic materials, has established their pre-eminent posi- 
tion among discharge lamps. Their increasing usefulness is further indicated by 
the wide range of design possible with various combinations of operating pressure 
and electrical loading, as well as by modification of spectral quality through the 
use of fluorescent material on the surrounding envelope of the low-pressure arc, 
or the introduction of additional metallic vapors in those of higher pressure. 

Only the last chapter of the book is devoted to applications of the sources, a 
treatment which might have been considerably expanded with great advantage to 
such groups as are represented by the readers of this JOURNAL. 

The author is to be complimented for including an excellent index as well as an 
Appendix of valuable supplementary information. 


General Electric Company 


Standards Recommendation 

On December 16, 1948, the Standards Committee recommended reapproval of 
the American Standard for cutting and perforating 35-mm negative raw stock, 

Reapproval or revision of this standard has been under consideration since 
October, 1946, when review of all the existing Z22 standards was undertaken by 
the ASA Sectional Committee, Z22. The long delay in determining what action 
should be taken was caused by the difference which exists between the 35-mm 
negative and positive perforations. This difference has caused considerable 
trouble in connection with printing 35-mm color release prints where extremely 
accurate registration is necessary. Consequently, the whole matter of 35-mm 
perforating was reinvestigated with the thought that perhaps a universal positive- 
negative perforation could be agreed upon. ' Such a perforation based upon the 
original proposal by Dubray and Howell is now in the process of development and 
a proposed standard will be published in the JOURNAL in the near future. 

The use of the negative perforation, however, has become so firmly established 
in the industry that elimination as a standard at this time does not seem possible. 
Therefore, the foregoing recommendation that it be reaffirmed as an American 
Standard has been made. Technically, the draft which is now being submitted 
to Sectional Committee Z22 is identical with the 1944 edition. The method of 
dimensioning, however, has been modified so as to be in accord with present in- 
dustrial practice. 

Journal Exchange 

Mr. G. D. Murphy wishes to dispose of a complete set of JOURNALS of the SMPE 
starting with the October, 1930, issue. Persons interested in purchasing these 
copies should write to him at R.F.D. 2, Rockville, Md. 


New Products 

Further information concerning the material described below can 
be obtained by writing direct to the manufacturers. As in the case 
of technical papers, publication of these news items does not consti- 
tute endorsement of the manufacturer's statements nor of his products. 


The Miles Reproducer Company, 812 

Broadway, New York 3, New York, re- 
cently announced their latest Film- 
graph Model RTD. This machine 
makes it possible to produce syn- 
chronized personal talking pictures 

directly on developed black-and-white 
or color 16-mm family motion pictures 
for permanent playback, without proc- 
essing or darkroom when used in con- 
junction with the standard 16-mm 
silent projector. 

Filmgraph employs a jewel stylus to 
indent permanently a fine sound groove 

0.002 inch wide and to reproduce it in- 
stantly through the same apparatus. 
The sound track is made at either edge, 
between the sprocket holes and the 
frames of the picture, on the glossy side 
of the film. Several sound tracks may 
be indented side by side, if desired, and 
will not show on the screen, since they 
are hidden by the frame of the aperture 
in the projector. 

To make a recording, the Filmgraph 
is placed in front of the projector in such 
a manner that it will not interfere with 
the picture on the screen. The film is 
threaded through the projector in the 
usual way except that the film is 
brought up to the reel of the recorder 
instead of the feeding reel (upper 
magazine) of the projector. Sufficient 
leader film is used to bring the sight and 
sound in synchronism. 

The new model is portable, measuring 
4 J /2 X 8 X 9 inches and weighs less 
than 10 pounds. A microphone and an 
external loudspeaker mounted on a 
baffle are supplied as part of the system. 
The loudspeaker is equipped with a 15- 
foot cable to permit placing it behind 
or near the screen. 

For synchronizing talking pictures in 
conjunction with 8-mm picture films 
Model MRC is recommended. This 
model has a recording capacity of ap- 
proximately one hour, using 100 sound 
tracks on each face of the safety film. 


New Products 

Further information concerning the material described below can 
be obtained by writing direct to the manufacturers. As in the case 
of technical papers, publication of these news items does not consti- 
tute endorsement of the manufacturer's statements nor of his products. 

Single-Case Filmosound 

Bell and Howell Company, 7100 
McCormick Road, Chicago 45, Illi- 
nois, announces the development of a 
new single-case, sound motion picture 
projector. The single-case filmosound, 
weighing only 43 3 /4 pounds, is designed 
with a 6-inch speaker mounted on a re- 
movable door in the side of the pro- 
jector case for carrying convenience. 
The door may be swung out at right 
angles to the case and the speaker 
operated from this position, or it may 
be removed from the case and placed 
near the screen. 

Speaker and projector are connected 
by a 40-foot cable, and up to 60 feet of 

additional cable may be added, if 
necessary. A 10- watt amplifier is pro- 
vided allowing the use of a larger 
speaker as an accessory, if desired. 


Television engineer and program-director. Belgian engineer, age 
38. Speaks English, French, German, Dutch. Graduated from 
Technical High School of Paris and Brussels as engineer and profes- 
sor in electronics and acoustics. Active Member SMPE. Has 
permit for immigration into United States. Highest qualifications 
and references. Excellent business experience and musical back- 

Three years, motion picture engineer, laboratories Thomson- 
Houston, Paris. 

Nine years, motion picture and television engineer, Philips' 
Physical Laboratories, Eindhoven. Director in charge of Philips' 
television caravan touring all over Europe. 

Three years, technical director, Decca Recording plant, Brussels. 
In charge of recording and production. Reliable and capable ex- 
ecutive for erection or direction of television station or television 
mobile units. Minimum salary wanted, $7500. Write, Engineer, 
Postbox 291, Antwerp, Belgium. 





Part I March, 1949, Journal Part II High-Speed Photography 




Foreword . 3 

John H. Waddell 

What Is High-Speed Photography? 5 

Maynard L. Sandell 

Electrical-Flash Photography 8 

Harold E. Edgerton 

New High-Speed Stroboscope for High-Speed Motion 

Pictures 24 

Kenneth J. Germeshausen 

Lamps for High-Speed Photography. . 35 

R. E. Farnham 

Motion Picture Equipment for Very High-Speed 

Photography 42 

Brian O'Brien and Gordon G. Milne 

Methods of Analyzing High-Speed Photographs 49 

Wade S. Nivison 

New Developments in X-Ray Motion Pictures 61 

C. M. Slack, L. F. Ehrke, C. T. Zavales, 
and D. C. Dickson 

High-Speed and Time-Lapse Photography in Industry and 

Research 71 

Henry M. Lester 

Use of High-Speed Photography in the Air Forces 81 

E. A. Andres , Sr. 

High-Speed Photography in the Automotive Industry 90 

Richard O. Painter 

Applications of High-Speed Photography 97 

Max Beard 

Control Unit for Operation of High-Speed Cameras 107 

L. L. Neidenberg 

Lenses for High-Speed Motion Picture Cameras 110 

Alan A. Cook 

High-Speed Photographic System Using Electronic Flash 

Lighting 116 

William T. Whelan 





Published monthly at Easton, Pa., by the Society of Motion 

Picture Engineers, Inc. Publication Office, 20th & Northampton Sts., Eaaton, Pa. General 
and Editorial Office, 342 Madison Ave., New York 17, N. Y. Entered as second-class matter 
January 15, 1930, at the Post Office at Easton, Pa., under the Act of March 3, 1879. 


The Committee on High-Speed Photography of the Society 
of Motion Picture Engineers was organized in January, 1948, 
to further existing knowledge in high-speed photography and 
to disseminate that lore. The Committee also undertook to 
sponsor the development of equipment to make high-speed 
photography more usable through portability and better 

In order that the photographic engineering profession be 
brought up to date on the advancement of the art, a sympo- 
sium was held in Washington, D. C., on October 29, 1948. At 
this symposium various high-speed cameras were described as 
well as techniques used by governmental and industrial 

The Committee has been striving to have the issuance of 
papers on high-speed photography concentrated so people who 
are interested in the subject will not have to peruse many pub- 
lications in order to find out what are the latest developments 
in the science of high-speed photography. This supplement to 
the Society's JOURNAL has all of the papers that were pre- 
sented in Washington with the exception of one. The paper 
on the design of rotating-prism-type cameras is being revised 
in the light of comments made by John Kudar which appeared 
in an earlier issue of the JOURNAL. 

It is felt that the papers contained herein form the founda- 
tion for high-speed photographic work which is being con- 
tinued at the first International Symposium on High-Speed 
Photography to be held at the Hotel Statler in New York on 
April 6, 1949. The subjects covered not only include high- 
speed motion picture photography but high-speed still photog- 
raphy as well. There is material enough for a third sym- 
posium if the reaction is as favorable to the second as it was to 
the original one held in Washington. 

The Chairman of the Committee wishes to express his 
appreciation to all members of the Committee and to those 
who have assisted the Committee in providing the active pro- 
gram as they have. If there is sufficient interest in the subject, 
plans can then be made with the Headquarters of the Society 
of Motion Picture Engineers to take care of the special needs 
of those interested in high-speed photography and photo- 
graphic engineering. 



High-Speed Photography Committee 

What Is High-Speed Photography P 




IN SPITE OF THE FACT that a great deal of work has been done in the 
field of high-speed photography, much of it has been isolated, and 
the research has not been centralized for the convenience of everyone. 
Your Society has realized this shortcoming, and early in 1948 proposed 
that a permanent Committee on High-Speed Photography should be 
made a part of its organization. It was decided that this activity 
should embrace both still and motion picture photography because 
of the close relationship between the two in this instance. 

In order to understand what is meant by high-speed photog- 
raphy, it is necessary to define it. In the field of still photography, 
with the knowledge and materials at hand, it is estimated that no con- 
ventional mechanical shutter is likely to be designed with a speed in 
excess of Viooo of a second. Any still picture made at a rate exceeding 
this will be considered a high-speed photograph. In the field of 
motion pictures, the mechanical limitations imposed on an intermit- 
tent mechanism will probably prevent its use at speeds in excess of 
250 frames a second. For this reason, your Committee has decided 
that motion pictures at a rate in excess of this will be considered 
high-speed motion pictures. 

High-speed still photography is not new. Nearly a hundred years 
ago, Henry Fox Talbot, who pioneered in so many phases of photog- 
raphy, successfully photographed a section of the London Times 
while it was being rapidly whirled on a disk. Talbot's light source was 
the spark generated by the discharge of a Ley den jar through an air 
gap. The spark's duration, probably less than a microsecond, effec- 
tively stopped the motion of the disk during exposure, permitting the 
finest print of the newspaper to be read when the plate was developed. 

From this humble beginning, the technique of high-speed still 
photography steadily improved with the increasing demand for its 
use in the science of ballistics. High-speed silhouettes, or shadow- 
graphs, became commonplace in the literature, and photographs, by 
the schlieren method, of sound waves and thermal disturbances in 
gases appeared. These pictures were made in a matter of millionths 


6 SANDELL March 

of a second. However, as an indication of how far the science has 
progressed, it might be well at this point to cite an interesting anomaly 
in this phase of photography. By using a Kerr cell as an electro- 
optical shutter, photographs have been made at a speed as high as 
4 X 10~ 9 second. To illustrate how incredibly short this period of 
time is, let us consider it hi terms of space. If the distance from 
Washington, D. C., to Hollywood, California, which is something less 
than 3000 miles, were represented as one second, then 4 X 10 ~ 9 
second would be only about seven tenths of an inch along the road 
from the Nation's Capital to the film capital. 

With the development of gaseous discharge tubes for electronic 
flash, high-speed still photography by reflected light in both black and 
white and color was made possible. What was once a laboratory 
curiosity became the everyday tool, not only of science and industry, 
but of news, commercial, and portrait photographers. 

High-speed motion pictures progressed more slowly because of the 
lag in developing cameras and projectors. In the early days, it was not 
convenient to produce a series of pictures in rapid sequence. The 
attempts of Muybridge required that a separate camera be used for 
each picture in the sequence. Furthermore, there was no satisfactory 
solution to the problem of viewing such pictures if they were pro- 
duced. The Zoetrope, or wheel of life as it was called, was a tem- 
porary expedient. It consisted of a shallow, slotted drum inside of 
which was mounted the strip of pictures to be viewed, the number of 
slots corresponding to the number of pictures. As this drum was 
rotated, a fleeting view of each picture was obtained in sequence as 
the slots passed by, and if the motion was rapid enough to overcome 
the eye's persistence of vision, an illusion of motion resulted. 

It was not long after the motion picture camera and projector were 
invented that attempts were made to alter the natural speed of sub- 
jects on the screen by changing the camera speed, the projector speed, 
or both. Action was generally accelerated to achieve ludicrous per- 
formances in comedies, while for better analysis of motion, as in 
sports events, the action was slowed down. Since projection speeds 
soon became more or less constant, these changes of tempo were ob- 
tained by varying the taking speed. 

As the value of slowing down action for analytical purposes became 
more evident, attempts were made to increase the taking rate, until 
speeds in excess of 150 frames a second were attained. This resulted 
hi slowing up action as much as ten times when the film was projected 


at 16 frames a second. Although this was adequate for analysis of 
most human or animal action, it became apparent that much higher 
taking speeds would be necessary for studying fast mechanical move- 
ments. To accomplish this, some means had to be found for sup- 
planting the conventional intermittent mechanism of the camera be- 
cause of its mechanical limitations. When this was achieved, high- 
speed motion picture photography was born. 

Motion picture cameras capable of speeds greater than 250 frames 
a second generally employ a rotating prism or mirror, a series of 
rotating lenses, or a stroboscopic light source instead of the usual 
shutter and intermittent mechanisms. In the latter case, the camera 
is equipped with a commutator to synchronize the flashing light with 
the film movement, and the camera is operated in a dark or semi- 
darkened room. Such cameras cannot be used for photographing 
self-luminous subjects. Cameras employing the rotating-prism 
principle usually bracket the range from 250 to 10,000 frames a second. 
Speeds in excess of 10,000 frames a second fall in the classification of 
ultrahigh-speed photography and are usually obtained in what are 
known as strip cameras. 

Of special interest among ultrahigh-speed motion picture cameras 
of the strip type is that developed by Brian O'Brien of the University 
of Rochester. Though the results obtained with this camera are some- 
what lacking from the standpoint of resolution, the camera is capable 
of making fifteen million pictures a second, but for a period of only 
1 Aoo of a second. If this camera could be operated for as long as a full 
second, the resulting sequence of pictures when projected continu- 
ously at the normal speed of 16 frames a second would last for nearly 
11 days. 

Despite the handicaps under which the makers of high-speed still 
and motion pictures worked, the art has grown steadily since its incep- 
tion. Perhaps the greatest growth has taken place during the war and 
the few years following it, until today this branch of photography has 
become a frequently used and invaluable tool of science and industry. 

Specialists in every field of high-speed photography, including 
X-ray, infrared, ultraviolet, and color, have been invited to participate 
in the activities of this new Committee of the Society. Representa- 
tives of industry and the universities as well as members of the Armed 
Forces are contributing from their experience and knowledge. The 
papers presented on the following pages will cover various methods 
employed in still and motion picture high-speed photography. 

Electrical-Flash Photography 



Summary Within the past ten years the use of electrically produced 
flashes of light for photography has become widespread. This type of light 
has some very important technical advantages which eventually will win 
it a place in everyday photography in addition to those applications where 
high speed is required. Modern electronic lighting equipment is particu- 
larly suited for color photography since the color of the light closely approxi- 
mates daylight, the color remains constant, and the quantity of light per flash 
can be closely controlled. Furthermore, the electronic system can produce 
large quantities of light for large-scale color photography in studios with re- 
sults that cannot be duplicated with any other known type of lighting 

The object of this paper is to outline the circuits and components that are now 
in common use. The principles of light production are given as well as meth- 
ods of calculating exposures, especially for color films. A method of meas- 
uring integrated incident light from flash sources and a meter for that pur- 
pose are also described. 


WHEN ELECTRICAL ENERGY from a capacitor is discharged into an 
open spark a visual flash of light of very short duration is pro- 
duced. Modern electrical flash systems (speedlights, stroboscopic 
lights, and others) use the same principle of the spark, except the 
efficiency of light production is greatly increased. The improvement 
resides in the use of a noble gas, such as argon, krypton, or xenon, 
which is constrained in a glass or quartz tube from which air has been 

Xenon is the preferred gas, even if the most expensive, because it 
has a high efficiency of conversion of the stored energy from the 
capacitors into visual and photographic light. Likewise, of the gases 
that can be used, xenon gas produces a spectral distribution of energy 
close to daylight and therefore the xenon flashtube is used as a source 
for photography with daylight color film. The spectra of Fig. 1 show 
a comparison between daylight and the xenon-filled flashtube. Note 
that the unfiltered xenon spectrum has an excess of radiation in the 
near-ultraviolet and blue- violet portion of the spectrum. A filter is 
used to reduce this radiation by the right amount when such sources 


10 EDGERTON March 

are used to expose daylight-type color emulsions. Different types of 
color films respond differently to the short flashes of light; therefore, 
a different filter is required for each type of film. 

The xenon flashtube has several very important properties that are 
of great interest to those who photograph in color. First of all, the 
light output can be controlled accurately since the electrical energy 
can be determined accurately. Second, the color quality is not 
materially changed with slight voltage changes. Third, the color 
quality of the light is not changed with the life of the tube. 

Those who have used tungsten sources for color photography will 
appreciate the above xenon-tube advantages since the tungsten-lamp 
output and color temperature are a function of voltage and age of the 
lamp. The xenon flashtube, unlike the tungsten lamp, is not subject 
to spectral changes during the life of the tube. In other words, the 
color distribution is the same for a new lamp and for one that has 
experienced thousands of flashes. This color constancy is a charac- 
teristic of the electronic system of producing light where a noble gas 
is excited electrically. The color of light from a tungsten lamp changes 
during life since the filament changes its dimensions due to evapora- 
tion. Furthermore, the evaporated filament material that is deposited 
on the bulb acts as a filter. 

One of the most important characteristics of a flashtube is its 
efficiency, that is, its ability to receive electrical energy from a ca- 
pacitor and convert much of that energy into a radiant form of which a 
part is light. Fig. 2 gives the efficiency of a typical small flashtube as 
a function of voltage and capacitance. The output of the flashtube 
under given conditions can be obtained from this diagram by reading 
the efficiency and calculating, as follows : 


Q = light output of flashtube in lumen-seconds 
C = capacitance in farads 
. E = initial voltage of capacitor in volts 
CE 2 /2 energy storage in watt-seconds 

n = efficiency of flashtube (function of E and C as given in Fig. 2), 
lumens per watt. 

The highest efficiency is attained in the region of the "damage 
limit." Since glass cannot be held accurately to dimensions, this 
damage-limit region will vary for individual tubes of the same type. 




Therefore, it is advisable not to operate too close to the border. The 
manufacturers of flashtubes rate the tubes according to the maximum 
energy at a specified voltage. 

For any given shape of flashtube, the efficiency with xenon-gas 
filling can be varied by changing the pressure. Below a few centimeters 
of pressure, efficiency is almost a direct linear function of pressure. 
The spectral characteristics of the light output for low pressure (about 
1 centimeter for the tube whose characteristics are shown in Fig. 2) 


Fig. 2 Efficiency curves for xenon-flash spiral source 13 /ie X 1 6 A inches as 
used in General Electric flashtubes FT-210, FT-214, and FT-220. 

are characterized by less continuum and stronger line spectra. For 
pressures above 5 centimeters the efficiency-versus-pressure relation- 
ship flattens out to give a constant efficiency. It is in this region that 
most flashtubes are designed. 


The operating limits for the tube whose characteristics are shown 
in Fig. 2 are boxed in by several factors. The left side is limited by 
the "unreliable-starting" region. On the other side is the "self-start" 
limit where the voltage is enough to start the discharge without the 
ionizing assistance of a high trigger potential on the outside of the 

12 EDGERTON March 

tube. The "unreliable-starting" and "self -starting" limits are in- 
fluenced by pressure. Both become higher with increased pressure 
gas. Slight impurities of gas likewise increase these two limits. The 
top limit is the "damage limit" where crazing of the tube is caused by 
the arc. 

Relay-operated flashtubes are designed so that the "self -starting" 
limit is below the voltage to which the capacitor is charged. This is 
accomplished for any particular design by reducing the pressure until 
the desired limit value of voltage is attained. 

Life of a flashtube is a function of the energy used per flash and the 
total number of flashes. At present, the life of the existing flashtubes 
is in excess of 10,000 flashes and thus seldom a problem. In fact, the 
flashtubes might be soldered or permanently wired into the reflector 
mountings since lamp replacement is a rare event. 

A flashtube reaches the end of its useful life for several reasons such 

(1) It becomes a "hard-starter" due to the release of contaminat- 

ing gases. 

(2) It becomes a "self-starter" due to the absorption of gas. 

(3) It becomes inefficient because of electrode sputtering and 

' absorption of light by the darkened walls. 

(4) It is broken physically or has a cracked seal. 

Of these four factors, the last is often the most important especially 
where the flash equipment is used on location. 

For single-flash work the designer can load up the tube so that it 
works hi the high-efficiency region. However, if the tube must be 
flashed at a rapid rate then average heating may be the limiting feature 
and lower energy inputs per flash must be used. For example, the 
tube whose characteristics are shown as Fig. 2 will operate at about 
15 watts for continuous operation. Thus, at one flash per second the 
allowable input is 15 watt-seconds, that is, 7.5 microfarads at 2000 
volts. At this loading, Fig. 2 shows that the efficiency is about half 
of tHe maximum that can be attained with the tube. At higher 
flashing rates the efficiency will be still lower. 

Tubes to be flashed repetitively should be cooled by a blast of air 
so that the average power can be increased. Also, the tubes for this 
purpose should be made of quartz since quartz is capable of with- 
standing higher temperatures than glass. 




Figs. 3 and 4 show the circuit elements used in two of the common 
types of electrical-flash systems. There must be a capacitor-charging 
system which is capable of supplying enough current to charge the 
capacitor in the time available between flashes. This time usually is 
5 or 10 seconds for single-flash equipment. For the very large flash 
units it can be 30 seconds or more if the power input is limited, for 
example, to 30 amperes peak from a 110-volt lighting circuit. For 
multiflash applications the time may need to be l / m of a second or 
less. The flashing rate of a given tube is limited both by the charging 
rate and by deionization of the tube as well as by tube temperature. 
Special circuits 1 " 3 are required when very high-frequency operation 
is desired. 

Fig. 3 Elementary circuit of the 
"electronic," "instantaneous," or 
"trigger" type of electrical-flash 

Fig. 4 Elementary circuit of a 
"relay "-operated flash tube. The lamp- 
operating relay is connected in parallel 
with the shutter solenoid. Adjustment 
of the relay time delay is made until 
light can be seen through the lens. 

The power to operate a flashtube at a given frequency of F cycles 
per second is approximately the following: 

power = 


For the majority of photographic problems, a flash unit is not re- 
quired to operate continuously. Therefore, the electrical designer can 
overload his circuit elements, such as transformers, but he needs to 
design for adequate current-delivering properties. Power supplies for 
electrical-flash equipment are evaluated according to (1) final voltage 
or open-circuit voltage, (2) short-circuit current to charge the capac- 
itor at the instant after the lamp flashes, and (3) time to charge a 
given capacitor to 90 per cent of the final charge. 

Actual power to the power system may be double the value given 




by the above equation due to losses in the power supply. The elec- 
trical efficiency is seldom of importance in design except in the case of 
portable equipment where the number of available flashes from a 
given battery are concerned, or where the power circuit is overloaded. 
The average current / to charge a capacitor C, in T seconds to a 
voltage E, is approximately 

I = -jr amperes. 

At the beginning of the charge cycle the peak current will need to be 
several tunes this value. 

Fig. 5 Oscillographic traces of the light output (top) and current (bottom) 
for three different discharge capacitors, 120, 38, and 10 microfarads. The 
peak light for the 120-microfarad trace is at the top and has a value of about 
27 X 10 6 lumens. The peak current is about 375 amperes. The above tests 
were made on the spiral as used in the FT-210, FT-220, and FT-214 (General 
Electric tubes) at 2000 volts. Timing dots are spaced 61.2 microseconds apart. 

The discharge current, Fig. 5, resembles that of a resistance- 
capacitance circuit in form, especially for the large values of capaci- 
tance. The peak current for the 120-microfarad example is about 
375 amperes. Assuming the capacitor initial voltage of 2000 holds, 
the apparent initial resistance of the flashtube is 2000/375 = 5.3 
ohms. The tune constant for this example is 120 X 10 ~ 6 X 5.3 = 
635 microseconds, which checks roughly with the time required for 
the current to decrease to 37 per cent of the peak. The light at this 
tune is 20 per cent of the peak. 

The main differences in commercial single-flash units are brought 
out by these questions: 

(1) How are the units supplied with power? Studio units are 


connected to ordinary 60-cycle power outlets but portable equipment 
must be battery-operated. 

(2) How much light output is produced by each unit? Eventually 
standards will be set up so that the lumen-second or beam-candle- 
power-second output of different tubes or tubes and reflectors can be 
measured. Likewise the beam-candle-power-second distribution of 
light from the reflectors as a function of angle will be evaluated since 
this is always an important factor. 

(3) How is the flash lamp caused to flash? How is synchroniza- 
tion of the flash with a camera shutter accomplished? Two types are 
in common use (a) the "trigger" type usually started instantaneously 
by an electronic trigger tube, and (b) the "relay" type whereby the 
tube is connected directly across the capacitor by the relay. Both 
methods will be discussed later in this article since this synchronizing 
question should be foremost in the mind of anyone who desires to con- 
nect a camera and flash unit together. 


The exposure tune for photographs taken with electrical flash 
lighting is determined entirely by the flash duration if the continuous 
light that is present does not also contribute to the exposure. The 
shutter time plays no part in determining the exposure time with 
flashtubes since the flash is always much shorter than the shutter-open 
interval. In fact, the function of the shutter, when using flashtubes, 
is to prevent extraneous continuous light from exposing the picture. 

Exposure time with electrical flash can be defined as the time that 
the light is above Vs of ^ ne peak light. When studies of duration, as 
in Fig. 5, are made with a specific flashtube, a curve such as Fig. 6 re- 
sults. In general the duration is shorter with high voltage and with a 
small capacitor. 

Flash duration or exposure time is seldom of importance except 
where rapidly moving objects are to be photographed or where close- 
up photographs of moderately fast motions are made. Any specific 
case can be checked for acceptable flash duration when the velocity of 
the object is known and when the minimum acceptable blur is ex- 
pressed. For example, a .22-caliber bullet with a velocity of 1100 feet 
per second requires an exposure of about 3 X 10 ~ 6 second (3 micro- 
seconds) if the maximum allowable blur motion is 0.04 of an inch. 
With a flash duration of 100 microseconds ( second) the bullet 
will move more than an inch. 




Special equipment is available for bullet photography with a 2- 
microsecond exposure time. A short flash is obtained by the use of 
a small flash capacitor ( l /$ microfarad) charged to high voltage (7000 
volts). The flashtube also is designed with a short discharge path 
with a fairly large cross section so that the tube resistance is low but 
still sufficient to prevent any oscillations of the discharge current. A 
small amount of hydrogen mixed with argon gas has been found to 
give some quenching of the afterglow in the gas after the current has 
ceased to flow. 

Fig. 6 Exposure time or flash duration as measured 
while light is above Vs of peak light as a function of 
capacitance and voltage for the flashtube as used in the 
General Electric tubes FT-210, FT-220, and FT-214. 
Data from General Electric Company. 

The very large energy discharges (3000 to 30,000 watt-seconds) into 
the quartz flashtubes as used for sources for large-scale color photog- 
raphy, would produce an annoying amount of noise if flashed without 
a small series inductance (about 5 millihenries). Such a series induct- 
ance does not materially increase the flash duration since its main 
function is to reduce the rapid rise of current at the start of the dis- 
charge. The series inductance also aids in deionization since it forces 
the current to flow longer than usual and thus brings the residual 
capacitor voltage to a lower value. 



The first problem after an electric-flash unit is obtained is how to 
connect it to a camera. Fortunately, many of the manufacturers of 
cameras and shutters are aware that the electric-flash system is due 
for growth and have been designing electrical contacts in their devices 
for flashing the light. These contacts need not be capable of carrying 
very much current when used with electronically triggered flashtubes. 

Unfortunately the synchronizing requirements are different for the 
various types of chemical and electrical flashbulbs that may be used 
with a shutter. Three synchronizing requirements meet most of 
the problems that arise. 

(1) Instantaneous contacts that close an electrical circuit at the 
exact instant that the shutter blades reach a full-open yosition as re- 
quired for triggered speedlights of the electronic types. Such contacts 
are called "X" types. 

(2) A 5-millisecond shutter delay. Here the lamp circuit is required 
to close a few milliseconds before the shutter reaches its open position 
so that the peak light output of the SM flashbulb at 5 milliseconds 
may occur while the shutter is open. 

(3) A 20-millisecond shutter delay. Here the lamp circuit closes 
15 to 17 milliseconds before the shutter reaches its open position. This 
delay is required for the average chemical flashbulb to reach its peak 
light output after the primer circuit is closed. 

The delay mechanisms are more difficult to manufacture and to 
adjust and keep in adjustment. For many years, in fact since the 
advent of the chemical flashbulb, various external magnet and relay 
devices for firing the chemical bulb and obtaining a delay have been 
hi widespread use. These devices as well as shutters including delay 
mechanisms will continue to be used to synchronize chemical flash- 
bulbs. It is interesting to note, however, that a camera with an in- 
stantaneous synchronizer "X" set of contacts can be used with any 
5- or 20-millisecond chemical bulb by slowing down the shutter speed 
to */25 second. Eventually all shutters and cameras should be supplied 
with instantaneous types of "X" synchronizing contacts since such 
devices are easily arranged hi shutters. 

However, electric-flash speedlight equipment of the electronically 
triggered ionization types cannot be flashed with a shutter delay since 
the light flash will be extinguished by the time the shutter opens. 

18 EDGERTON March 

The "relay" type of electrical-flash equipment does have a time de- 
lay that is usually adjustable by the user. The delay is a function of 
battery voltage, temperature, position, circuit resistance, and so 
forth. When the "relay" system is used, two delay mechanisms must 
function reliably in order for the light to occur when the shutter is 
open. With long shutter tunes, this can be done. However, dif- 
ficulty may arise at fast shutter speeds if either the shutter delay or the 
firing relay exhibits any time variability. Unlike the chemical flash- 
bulb the electrical flash is short and a blank will result if the timing 
is off. 

Fig. 7 Method of calculating the guide factor 
for front-lighted photography. 

The current required to operate the relay in relay-fired flash units 
may be too high for the internal contacts in some shutters. Either 
the relay will not operate or the contacts will be damaged by the 
excessive current. 


The photographer who obtains an electrical-flash unit is mainly in- 
terested in the guide factor, which is the product of the lamp-to-subject 
distance and the camera aperture (see Fig. 7), for a suitable photo- 
graphic result. Guide factors for flash units are experimentally ob- 
tained by trial and error. An equation is given below which shows 
how the various factors influence the guide factor for the simple light- 
on-the-camera type of lighting. 

D/ = \/KMQ = guide factor 



D the lamp- to-subject distance in feet 
/ = the camera aperture 

= the lamp output in lumen-seconds 

= the reflector factor, a number that gives the ratio of light at the center 
of the beam with the reflector to the light without the reflector. Com- 
mercial reflectors often have an M factor of from 4 to 12. 

K an experimentally determined constant that corresponds to conditions 
suitable for satisfactory photography with any particular type of 
emulsion. See Table I for approximate values of K. 



Film Type 


Incident Exposure at 
Subject for //3.5 


Kodachrome 35-mm 





Kodachrome professional 




cut-film type 





Ansco Color tungsten 






Super Panchro Press 







* Value used depends upon processing and type of negative desired. 

Actually it is not as simple as it appears from above since photo- 
graphs are very seldom taken with front lighting of the subject. 
Therefore, the actual guide factor for any given equipment and condi- 
tion should be found by direct experiment after preliminary estimates 
have been made using the theory that has just been presented. 


Whenever one carries portable flash electric equipment on an 
assignment the question is always asked, "Why not make it lighter?" 

For portable work every effort should be made by the user to get 
the maximum out of his film-processing and lens equipment so that 
he can get his assignment with a minimum of light. 

With the minimum light output specified, it is then up to the equip- 
ment designer to get a flashtube of the greatest efficiency so that his 

20 EDGERTON March 

watt-second storage capacity will be as small as possible. Watt- 
second storage capacity means weight and for this reason the watt- 
second value should be as small as possible. 

A very considerable saving of weight could be obtained by over- 
volting the electrical capacitors that serve as storage devices for the 
required watt-seconds. For example, doubling the voltage would 
result hi a saving of weight by a factor of 4. However, the capacitor 
may break down, and probably will, since such flash capacitors are 
already rated as high as possible. It then comes to a decision of what 
chances one can take on capacitor failure. Many of the designers of 
flash equipment have erred on the side of overstress and capacitor 
failures have been the result. 

For professional flash equipment in a studio, weight is of no con- 
sequence compared to the inconvenience caused by capacitor failure. 
It behooves the user of large studio equipment to insist upon conser- 
vatively designed capacitors even at the cost of weight and price in 
order to enjoy failure-free performance. 

The weight of capacitors of a particular voltage rating is almost a 
direct function of the energy storage. From the equation for the 
guide factor, the weight is then proportional to the square of the 
guide factor, considering the case where the lamp efficiency and the 
reflector factor are constant. The above rule enables a flash-circuit 
designer to estimate the weight of a proposed flash unit in terms of an 
existing unit. 

For example, a 20-pound flash unit has been found to have a guide 
factor of 20 when used with color film. Suppose a guide factor of 200 
is desired, that is, a tenfold increase. The weight of the new flash 
unit will be ten squared or 100 times that of the first. A 2000-pound 
unit is indicated. 


The light meter for integrating light, which has been described more 
completely elsewhere, 4 ' 5 uses a vacuum photoelectric tube to produce 
a current that is proportional to instantaneous incident light. This 
current is integrated against time in an electrical capacitor and an 
electronic voltmeter measures the integrated value as voltage. The 
voltage is proportional therefore to the integral of the incident light 
against time, which is exposure. 

An incident-light meter that measures foot-candle-seconds (lumen- 
seconds per square foot) has two main uses. The first is to measure 




the lighting level at the subject for exposure determination, especially 
with color emulsions. For any given material and camera aperture 
the time-light product at the subject needs to be of a specific value. 
The lighting is then arranged until the lighting level is correct. A 
further refinement of this system is to use the meter to measure the 
light received on the ground glass from a white target card at the sub- 
ject. The advantage of this system is that all factors such as lens 

Fig. 8 Exposure meter which reads incident light in foot-candle-seconds 
(lumen-seconds per so.uare foot) from a xenon electronic-flash source. The 
meter also accepts a photocell probe for use on the ground glass of a camera. 
(See Fig. 9.) 

absorption and bellows extension are included. With this arrange- 
ment there is only one correct reading regardless of aperture for any 
specific type of film. 

The second use of the meter is to measure the output of flashtubes 
and of the tubes in reflector systems. The horizontal candle-power- 
second output of a lamp is obtained by multiplying the incident 
lumen-seconds per square foot by the lamp-meter distance squared. 
Beam candle-power-seconds from a tube in a reflector is calculated 
in the same manner where the distance is greater than about ten re- 
flector diameters. 

The beam candle-power-seconds (b.cp.s.) is approximately equal 





b.cp.s. = M 

for the average flashtube. The spherical distribution from the lamp 
is not uniform in space, resulting in a factor of about 10 instead of 
4ir for the ratio of the total lumen output to horizontal candle-power 


The various photographic problems that seek solution by means of 
electric-flash lighting can be roughly classified as follows. Those that 

Fig. 9 Probe measurements of exposure on 
the ground glass is illustrated above. The 
photocell in the probe measures the light imaged 
from a white card at the subject. 

(1) Large quantities of light of consistent color quantity and 
quality for professional studio color photography. Units of up to 
20,000 watt-second capacity are now in use. 

(2) Short flashes of light for rapidly moving objects such as bullets. 
Equipment with 2-microsecond flashes is in active use. A shorter 
flash with adequate light output appears to be difficult to obtain due 
to afterglow in the gas. 


(3) A series of pictures at a rapid rate for studying objects that 
move quickly. A companion paper by K. J. Germeshausen* describes 
the hydrogen-thyratron modulator high-speed stroboscope light 
which seems to have great promise as a pulsed light source. 

(4) Monochromatic light for use in interferometers. A short flash 
is also desired. 

(5) Small intense source for silhouette and microscope photog- 

In conclusion, it should be pointed out that the electric-flash system 
is a most powerful method for both everyday photography and 
difficult scientific research problems. Many new applications and 
uses will undoubtedly result in the next few years. 

* See pages 24-35. 


(1) F. E. Carlson and D. A. Pritchard, "The characteristics and application 
of flashtubes," Ilium. Eng., vol. 42; February, 1947. 

. (2) H. E. Edgerton, K. J. Germeshausen, and H. E. Grier, "Multiflash photog- 
raphy," Photo Technique, vol. 1, p. 14; October, 1939. 

(3) H. M. Lester, "Electronic flashtube illumination for specialized motion 
picture photography," /. Soc. Mot. Pict. Eng., vol. 50, pp. 208-233; March, 1948. 

(4) H. E. Edgerton, "Photographic use of electrical discharge flashtubes," /. 
Opt. Soc. Amer., vol. 36, pp. 390-399; July, 1946. 

(5) H. E. Edgerton, "Light meter for electric flash lights," Electronics, p. 78; 
June, 1948. 

(6) John S. Carrol, "Principles and design factors of electronic photo-flash 
units," Elec. Manufacturing, vol. 39, p. 47; April, 1947. 

New High-Speed Stroboscope for 
High-Speed Motion Pictures 



Summary This paper describes a high-speed stroboscopic light for use 
with high-speed motion picture cameras. It provides flashing rates up to 
7000 per second and an effective flash duration of 1.5 microseconds. Ex- 
perimental results are described showing the improvement in picture quality 
obtained when the light is used in conjunction with a Fastax camera. 


increasing use in the solution of industrial and scientific prob- 
lems. The need for a tool to analyze motions too rapid for the eye to' 
follow is readily apparent, especially one that does not require the 
attachment of measuring devices to the object under study. A num- 
ber of high-speed motion picture cameras have been developed, all of 
which employ a continuously moving film and in general two methods 
are employed to stop the motion of the image with respect to the 
film. The first of these methods is primarily an optical one involving 
the use of rotating lenses, prisms, or mirrors; the second is the use of 
stroboscopic light to illuminate the subject, the flashes of light being 
of such short duration that the film does not move appreciably during 
the exposure. 

Typical cameras of the first type are the Eastman high-speed 
camera and the Western Electric Fastax. Both of these cameras em- 
ploy a rotating prism to compensate for the continuously moving 
film. An example of the second type is the General Radio high-speed 
motion picture assembly. 

Each of these two types of high-speed cameras has certain ad- 
vantages not possessed by the other. The optical types are generally 
smaller and simpler to operate. They may be used to photograph 
self-luminous subjects and under some conditions pictures may be 
taken outdoors in bright sunlight without additional sources of illu- 
mination. The stroboscopic type has the outstanding advantage of a 
very short exposure tune, on the order of 2 to 10 microseconds, which 
24 . 


prevents blur when photographing rapidly moving objects. Further- 
more, stroboscopic illumination is relatively cool, a distinct advantage 
when photographing subjects liable to damage by heat. 

Stroboscopic illumination is limited to small areas, not more than 4 
to 8 square feet for a single lamp, and cannot be used in the presence of 
strong general illumination. However, the majority of subjects re- 
quiring short exposure times are small and can be illuminated with a 
single lamp. Optical-type cameras are limited in definition, partly 
because of motion of the subject and partly because of distortions in- 
troduced by the moving optical system. 

An ideal combination would be an optical-type camera that could 
be synchronized readily with stroboscopic light. For self-luminous 
subjects or large subjects particularly in sunlight, the camera could 
be used as is. For subjects requiring short exposure time, or where 
the intense heat associated with continuous light is objectionable, the 
same camera could be used with stroboscopic light, giving improved 
definition in the picture. 

In the past, stroboscopic lighting has been criticized as bulky and 
complicated, often requiring the services of an expert to operate it, 
and, in addition, limited in maximum operating frequency. Modern 
designs of equipment overcome these objections to a large extent. 


The new high-speed stroboscope represents a further development 
of equipment designed during the last war for the California Institute 
of Technology. Their problem required the illumination of a large water 
tank for studies of underwater projectiles. Stroboscopic lighting was 
indicated because of the short exposure required but none of the 
equipment then in existence would meet their requirements of light 
output, flashing rate, and duration of exposure. 

The development of hydrogen thyratrons and their application to 
radar modulators 1 offered interesting possibilities as a solution to the 
above problem if a lamp of suitable characteristcis could be designed 
to match the modulator. Basically the circuit of the radar modulator 
as applied to the stroboscope is as shown in Fig. 1. 

A capacitor C is charged from a direct-current supply through an 
inductance LI, a rectifier T\, and an inductance Z/2. In this type of 
circuit the voltage on C will rise to nearly twice the supply voltage. 
Once the capacitor is charged to this voltage it will remain charged, 
since the rectifier T\ prevents it from discharging back into the supply 


and thus returning to supply voltage. If the impedance of L* is low 
compared to the impedance of LI, very little voltage will be developed 
across L 2 during the charging cycle. 

When it is desired to flash the lamp the hydrogen thyratron T 2 is 
triggered. This effectively connects the charged capacitor across the 
lamp and the inductance L*. If the voltage on C is sufficient to cause 
breakdown of the lamp, the capacitor will then discharge through the 
lamp causing a flash of light. In order that most of the energy in the 
capacitor be discharged into the lamp, the impedance of the lamp 
must be low compared to the impedance of L*. During the discharge 
of the capacitor and while the thyratron T^ is conducting, the rela- 
tively high impedance of LI prevents any appreciable current flowing 

from the supply through 2V 
Immediately after the capacitor 

PC becomes discharged the thy- 
LA np ratron T 2 extinguishes and thus 
permits C to become charged 

again in the manner previously 
Fig. 1 Basic circuit for high-speed described 

The ability of the hydrogen 

thyratron, as compared to other thyratrons, to carry heavy short- 
duration discharge current and to extinguish very quickly, make the 
circuit of Fig. 1 feasible. 

One of the major difficulties with the older type of high-speed 
stroboscope circuits 2 was the failure of the discharge circuit to become 
nonconducting or to deionize rapidly enough. This failure to deionize 
prevented recharging of the capacitor C at very short intervals and 
thus limited the maximum flashing rate of the stroboscope. The 
ability of the hydrogen thyratron to deionize rapidly permits much 
higher flashing rates, up to as high as 10,000 per second in properly 
designed circuits. 


To secure maximum power output from the hydrogen thyratron the 
load, or the lamp, should have 30 to 50 ohms impedance. Further- 
more,, for a specific circuit, proper operation will be secured only if this 
impedance is held to rather close limits. 

From the work of earlier experiments 3 it appeared that the im- 
pedance of flashlamps was generally much lower than the required 30 
to 50 ohms. It also appeared that lamp impedance was not a constant 
but depended upon the length of the discharge path, the diameter 
of the tube, gas pressure, and the value of the discharge capacitance. 




An investigation of these various parameters was made, extending 
them outside the range explored by previous experimenters. As a re- 
sult of these experiments a lamp design evolved having an impedance 
of approximately 30 ohms with the values of discharge capacitance em- 
ployed in the new stroboscope. This lamp has 
an arc length of approximately 4 inches. By 
means of a new type of construction, this 4- 
inch arc length is coiled into a cylinder 3 /s by 
3 /s of an inch, giving a relatively small source 
size that may [be used with high optical effi- 
ciency. A photograph of one of the lamps is 
shown in Fig. 2. 

A second problem in lamp design for high 
flashing rates is one of efficiency. Flashlamp 
efficiency is a function of energy per flash and 
in general, to obtain maximum efficiency, the 
energy per flash should be high. If the loading 
per flash is high, however, then at high flashing 
rates the average power into the lamp becomes 
excessive, since average power is the product of 
energy per flash and flashing rate. 

It was found by experiment that it was im- 
possible to employ optimum loading per flash 
without destruction of the lamp by overheat- 
ing. To maintain the loading per flash at the 
highest possible level two steps were taken. 
One, the lamp is made of quartz which will 
withstand much higher operating temperatures 
than glass, and two, the operating time of the 
lamp is limited to the. minimum needed to ex- 
pose a roll of film. 

For a 100-foot roll of 16-mm film at 4000 pictures per second, the 
total duration of the film is one second. With the loading per flash 
used, operating at the above speed for more than 2.0 seconds will re- 
sult hi overheating of the lamp. 


The general circuit of the stroboscope is shown in Fig. 3. Com- 
ponents to the left of the dotted line comprise the power supply and 


Fig. 2 High - speed 
stroboscope lamp. 




components to the right of the dotted line comprise the modulator 
portion of the stroboscope. As constructed, these sets of components 
are on separate rack and panel chassis but may be mounted together 
in a single cabinet. The complete unit (Figs. 4 and 5) mounted in a 

cabinet measures 22 inches wide, 15 inches deep, and 28 inches tall 
and weighs 180 pounds. 

The power supply is a conventional bridge-circuit rectifier with 
certain added control features. It supplies up to 500 milliamperes 




direct current at 4000 volts to the modulator panel. To maintain 
minimum size and weight all components are specially designed for 
this particular application, keeping in mind that the load on the supply 
is intermittent. 

Input power is normally 220 volts, single-phase, 60 cycles, at 15 
amperes maximum. By means of a built-in Variac the voltage on the 
primary of the power transformer can be held at 220 volts over a 
range of line voltages of 195 to 250 volts. Considerable thought was 
given to the choice of input voltage; 220 volts was decided upon be- 

Fig. 4 Front view of high-speed stroboscope with ac- 
cessory cables and the lamp. 

cause the power drain was too great for most 115-volt lines. Relay 
82 in the primary of the power transformer can be remotely controlled 
by means of a switch connected to the remote-control terminals. On 
this same relay are a set of contacts which may be used to control 
camera motors simultaneously with the stroboscopic light. 

In series with the remote-control circuit is a tuner switch $ 4 which 
may be adjusted to shut off the power after an interval of from 1 to 15 
seconds. An overload relay Si, also in the primary of the power 
transformer, shuts off the power instantaneously if excessive direct 
current is drawn from the power supply. 




The lamp circuit is essentially the same as shown in Fig. 1 with the 
addition of adjustable discharge capacitances. Three values are pro- 
vided, 0.01, 0.02, and 0.05 microfarad. Because of the voltage dou- 
bling obtained, as previously explained, the capacitors Ci, 2, and C$ are 
charged to 8000 volts. The choice of discharge capacitance depends 
upon the flashing rate. In Table I are listed the energy storage per 
flash for each capacitor, maximum flashing rate, and average power 
delivered to the lamp at that rate. 

Fig. 5 Rear view of high-speed stroboscope showing 
method of construction. The upper chassis is the modu- 
lator portion of the unit and the lower chassis the power 

A coaxial cable connects the lamp to the modulator. As much as 
75 feet of cable may be used without materially affecting the light out- 
put or duration of the flash. 

Voltage pulses to trigger the thyratron are obtained from an am- 
plifier as shown in Fig. 3. The amplifier may be driven in several 
ways. A contactor or switch such as is used on the General Radio 
camera can be connected to the contactor terminals. The light will 
flash when contact is made. A sine- wave voltage of approximately 
10 volts applied to the oscillator terminals will control the flashing 




rate over a range of 50 to 7000 flashes per second. If a steeply rising 
voltage pulse is applied to the input of the amplifier less voltage is re- 
quired to trigger the thyratron. For sharp pulses approximately 0.5 
volt at the input of the amplifier is adequate for satisfactory operation. 


Capacitor Value, 

Watt Seconds 
per Flash 

Flashing Rate 
per Second 






To synchronize the flashing of 
the lamp with the rotation of the 
prism in cameras such as the 
Eastman and Fastax, a magnetic 
or reluctance pickup has been 
developed. This device consists 
of a small, permanently magnet- 
ized iron armature mounted in 
a coil of wire. The armature 
is mounted so the teeth of the 
film sprocket pass near one end. 
Each time a tooth goes by the 
armature a voltage pulse is de- 
veloped in the coil. The pulse 



01 z 3 4- s~ 

Fig. 6 Light and current versus time 
for stroboscopic lamp. 

developed is of sufficient amplitude when applied to the pickup ter- 
minals of the amplifier to control the firing of the thyratron. 

Curves of light output and lamp current versus time are shown in 
Fig. 6. These curves are for a value of discharge capacitance of 0.02 
microfarad. As can be seen from the curves the duration of the light 
for the above capacitance is 1.5 microseconds to Vs peak light and the 
total duration is 4 microseconds. Characteristically the light persists 
an appreciable time after the current has stopped. This persistence 
or afterglow is held to a minimum by the addition of hydrogen to the 
gas filling. 

The total light output and the duration will depend upon the choice 
of discharge capacitance. In general the duration is approximately 
proportional to the capacitance; however, due to the change in lamp 


efficiency with loading, relatively more light is obtained with the larger 
values of capacitance. 

Precision of firing of the lamp with respect to the triggering means 
is very good. From the tune a signal is applied to the grid of the thy- 
ratron there is a delay of approximately 0.5 microsecond to the 
flashing of the lamp. The variation in flashing of the lamp with re- 
spect to the triggering signal is less than 0.1 microsecond. This 
same precision applies to signals fed through the amplifier but the 
over-all delay in firing is somewhat greater. 

Light output hi terms of effective exposure on the film depends on 
a number of factors such as the type of film, reflectors, and discharge 
capacitor. With a 5-inch diameter, diffuse aluminum reflector, ade- 
quate exposure is obtained at lens apertures of //4 on areas 1.5 feet 
by 1.5 feet, using 16-mm super XX film and a discharge capacitance 
of 0.02 microfarad. Larger reflectors, or specularly finished reflectors, 
will increase the light on the subject. For small subjects, where the 
light can be focused on a small area, it is possible to work at apertures 

Light output is independent of flashing rate and depends only on the 
value of the discharge capacitance. 


The high-speed stroboscope was used in conjunction with a Western 
Electric 16-mm Fastax to determine what improvement could be ob- 
tained by the use of stroboscopic light in contrast to incandescent 
illumination. Synchronization of the light with the rotating prism 
was accomplished by means of the magnetic pickup previously 

It was found that the position of the pickup had to be set very care- 
fully to insure that the prism was vertical when the light flashed. If 
this was not done a serious loss in definition resulted. Placement of 
the pickup was checked by observing the reflection of the image of the 
flashlamp. The flashlamp was placed on the axis of the lens-prism as- 
sembly and the position of the pickup adjusted until the image of the 
light was reflected back along this same axis. 

Figs. 7A and B show the improvement in definition obtained by 
eliminating blurring due to motion of the subject. Both are pictures 
of a 10-inch disk rotating at 7200 revolutions per minute, the camera 
speed in each case was 3200 frames per second, and the lens aperture 
//4. Fig. 7A was taken with incandescent light and Fig. 7B with 




stroboscopic light. From these two pictures it can be seen that for 
rapidly moving subjects stroboscopic light offers a considerable im- 
provement in definition. 

It should be pointed out that the sharp images obtained with strobo- 
scopic light are most advantageous where it is desired to make a 
frame-by-frame inspection of the film. For a purely qualitative 
analysis of films by projection, blurring due to motion may not be 
objectionable. If the pictures are taken at a sufficiently high rate to 
give a good slow-motion effect, at least 100 frames per cycle of the 
event, then blurring due to motion of the subject is not too noticeable 
during projection and may even aid the appearance of the picture. 

A With incandescent light. 

B With stroboscopic light. 

Fig. 7 Enlargements from 16-mm high-speed motion pictures of a 10-inch 
disk rotating at 7200 revolutions per minute. Taken at 3200 pictures per second. 

Figs. 8 A and B are photographs of National Bureau of Standards 
25X resolution charts. Fig. 8A was taken with incandescent light 
and Fig. 8B with stroboscopic light, the lens aperture in both cases 
being //4. These pictures are enlargements of a portion of the 16-mm 
frame in order to show the resolution obtained; the numbers on the 
pictures correspond to resolution in lines per millimeter. 

As can be seen from these two pictures, resolution is somewhat im- 
proved by the use of stroboscopic light. In Fig. 8B, 40 lines per 
millimeter can be resolved as against 28 lines per millimeter in Fig. 8A. 
This improvement in definition results in a better over-all appearance 
of pictures used for qualitative analysis and, again, is of considerable 
advantage where frame-by-frame inspection is to be made. 



In conclusion it may be stated that the use of stroboscopic light 
with rotating prism-type high-speed cameras such as the Fastax will 
result in better definition. The improvement is of particular advan- 
tage where frame-by-frame inspection of the film is to be made for 
purposes of analysis by measurement. Stroboscopic light will stop 
the motion of rapidly moving subjects and permit accurate measure- 
ments to be made of form, velocity, and acceleration. A further ad- 
vantage is gained in those cases where the subject is liable to damage 
by heat because of the relative coolness of the stroboscopic light. 

A With incandescent light. 

B With stroboscopic light. 

Fig. 8 Enlargements for 16-mm high-speed motion pictures of a resolution 


For many problems the results obtained with incandescent light are 
adequate and the improvement obtained by the use of stroboscopic 
light may not justify the additional complexity and cost of the lighting 


(r)" The Hydrogen Thyratron, Sect. 8.11 Pulse Generators, M.I.T. Radiation 
Laboratory Series, 1st ed., McGraw-Hill Book Company, New York, N. Y., 1948. 

(2) H. E. Edgerton, K. J. Germeshausen, and H. E. Grier, "High-speed photo- 
graphic methods of measurement," J. Appl. Phys,, vol. 8, pp. 2-9; January, 

(3) P. M. Murphy and H. E. Edgerton, "Electrical characteristics of strobo- 
scopic flash lamps," J. Appl. Phys., vol. 12, pp. 848-855; December, 1941. 

Lamps for High-Speed Photography 



Summary This paper first discusses in general the requirements of light 
sources for high-speed motion picture photography, and in detail, five types of 
light sources that might be used for this work. Included with the discussion 
of each of these sources, specific lamp types are recommended and the de- 
tailed characteristics are given. A new type of reflector bulb lamp, developed 
specifically for high-speed motion picture photography, is explained. 


As A RESULT of considerable discussion by the subcommittee on 
illuminants for high-speed photography, the following require- 
ments for an illuminant have been established. 

1. It should cover an area to be photographed of 4 X 4 inches. 

2. The minimum distance from the source to the area is 18 inches. 

3. The variation in the illumination from the center of the area to 
the corners should not exceed 2 to 1. 

4. It should produce a minimum value of illumination at the 
center of the target of 50,000 foot-candles, although 100,000 foot- 
candles may be required. 

5. A color temperature of the source of approximately 3500 degrees 
Kelvin is desirable. 

6. Lamps would be operated at full power for approximately 
10-second periods. 

Other desirable features include the following: 

1. To permit the light source to be kept close to lens-subject line 
it should be as compact as possible. 

2. The required illumination levels should be obtained with prefer- 
ably two lamps. 

3. In the case of incandescent lamps, the use of tungsten cleaning 
powder is suggested to minimize the effect of bulb blackening. 

Upon converting the 18-inch minimum t lamp-to-subject distance 
and the 4- X 4-inch area to light-source-distribution requirements, we 
find approximately an 18-degree beam spread. This is based on the 
beam candle power dropping to 50 per cent of maximum, 9 degrees 
each side of the center of the distribution pattern. 


36 FARNHAM March 


The remainder of this paper discusses the advantages and dis- 
advantages of available illuminants. Among those considered are: 
Incandescent lamps (with separate reflector and reflector-bulb 


Electrical-discharge lamps: mercury, fluorescent, and flash- 
Combustion sources. 

The Incandescent Lamp 

Without doubt this type of illuminant has been most widely used 
thus far and when its advantages and disadvantages are compared 
with those of other sources, it stands pre-eminent. 

Incandescent lamps can be manufactured in sizes from a few watts 
to 10,000 or more watts. They can be operated from storage bat- 
teries or standard lighting circuits of almost any voltage, without 
auxiliary control or regulating equipment. They operate equally 
well on alternating or direct current. They light and extinguish 
quickly, and can be designed for a wide range of lives consistent with 
the requisite color temperature or efficiency. The light source itself 
can be made quite concentrated, thus producing high brightness and 
permitting accurate control of the light by means of reflectors. The 
spectral-energy distribution of the incandescent source is continuous 
and makes possible satisfactory color pictures with materials adapted 
to the tungsten radiation. The variation of the light occurring with 
alternating-current operation is less than 5 per cent for lamps above 
100 watts on 60-cycle circuits. 

Among its disadvantages are the high temperatures that accompany 
high levels of illumination. While heat-absorbing filters can be used, 
they are usually bulky, and where high wattages are employed, circu- 
lating, cooling water with its complications, generally is necessary. 
It is more often the practice to operate without filters and ignore the 
heat. A stream of air on the area being photographed will help 

Compared to other illuminants the incandescent lamp is not the 
most efficient in terms of photographic effectiveness versus input 
watts. On the other hand, the light of tungsten-filament lamps can 
be efficiently directed on to the area being photographed, thereby 
often more than compensating for the lesser actinicity of the light. 

The wattages generally required for high-speed motion picture 




photography (at least 1000 to 1500) are such as to necessitate ade- 
quate power facilities, either in the form of ample-size wire or suffi- 
cient storage batteries plus heavy conductors, to insure that the 
lamps operate at their correct voltage. The photographic effective- 
ness of the light from tungsten-filament lamps is quite sensitive to 
variations in the applied voltage. 

Incandescent lamps applicable to high-speed photography may 
be listed in two groups, those provided with reflectors integral with 
the bulb and those requiring external reflectors. (See Table I.) 


Item Watts Bulb 

Hours Degrees Spread Max. Beam Ordering 
Life to 50% Max. Candle Power Code 

External Reflector Types 

1 750 T-12 25 

2 1000 T-12 10 

3 1200 T-12 10 

4 1500 T-12 25 


Integral Reflector Types 




































Items 1 to 4 are provided with medium prefocus bases and projection- type bi- 
plane filament construction. 

Items 5 to 9 are provided with medium screw bases. 
All lamps available in 115, 120, and 125 volts rating. 

Several of the lamps listed in Table I do not meet all the require- 
ments outlined at the beginning of this paper. They are lamps 
that have been used in high-speed motion picture photography, plus 
one new type. Items 1, 2, 3 are motion picture projection lamps hav- 
ing highly concentrated light sources and operating at high efficiency; 
they must be used in conjunction with a reflector. Item 4 is a lamp 
developed a number of years ago for high-speed photography and is 
intended to be used with a deep-bowl polished-aluminum reflector of 
an elliptical contour. 

38 FARNHAM March 

Items 5 and 6 are standard projector and reflector spotlamps used 
in many applications where a lamp of this type is desirable. They 
are mentioned simply because some photographers have taken lamps 
of 115 volts rating and operated them on 220 volts for short periods of 
time, thus obtaining a severalfold increase in light output. Item 7 is 
the reflector floodlamp of the photoflood type and is included in the 
table because it is a more convenient lamp where the area to be photo- 
graphed is relatively large. Obviously, quite a number of lamps 
would be required to obtain the requisite intensity of illumination. 
Item 8 is the reflector photospot lamp and is one of the most used 
lamps for high-speed motion picture photography. 

Item 9 refers to a new lamp developed especially for high-speed 
motion picture photography, keeping in mind the specifications dis- 
cussed at the start of this paper. Two of these lamps will produce 
65,000 foot-candles on an area 4X4 inches at a distance of 18 inches. 
Naturally, experience with this lamp is limited but it appears to be 
capable of producing quite good results. It is intended for intermit- 
tent burning, and for best performance the operating periods should 
be no longer than 15 or 20 seconds at full power. 


Mercury lamps, with the exception of the H6, have such funda- 
mental disadvantages that they can seldom be used. Nevertheless, 
it is well to include a brief discussion of this source. 

The advantages of mercury lamps, while few in number, are as 
follows: The color of the light is good photographically where black- 
and-white materials are involved but it is not suited for color pho- 
tography. Likewise, on the basis of equal photographic effectiveness, 
there is considerably less heat. 

The light output follows the cyclic variation of an alternating cur- 
rent quite closely, with the result that there would be a periodic series 
of underexposed pictures interspersed with overexposed pictures; 
consequently, it is necessary in using mercury lamps for this applica- 
tion "to operate them from direct-current sources. They require 
auxiliary ballast equipment whether operated from direct- or alternat- 
ing-current circuits, which becomes an important factor in the event 
that the lighting equipment is transported to the job. The majority 
of them require from four to five minutes to come up to full bright- 
ness and an even longer period in the event that the light has been 


turned off. The source brightness of mercury lamps ranges from 
approximately 175 candles per square inch to about 6000 candles per 
square inch for the commercially available types. This might be 
compared with the source brightness of about 20,000 to 24,000 
candles per square inch for concentrated-filament incandescent 
lamps mentioned in items 1 to 4 of Table I. Thus, it is much more diffi- 
cult to obtain high illumination levels with these more extended sources. 
The one exception mentioned above, namely the H6 lamp, has a 
source brightness of approximately 200,000 candles per square inch, 
more than 10 times that of the incandescent source. Furthermore, 
the lamp has a light source of about Vie of an inch in diameter and 1 
inch long, thus making it possible with suitable reflecting equipment 
to obtain enormously high illumination levels. One of the types 
of the H6 lamp, the A-H6, operates in a water jacket which removes a 
large part of the infrared radiation and thus a comparatively cool 
light can be obtained, an important factor in certain types of high- 
speed photography. The lamp comes up to full brilliancy within 1 or 
2 seconds, which removes an objection characteristic of the other types. 
However, its light follows the cyclic variation of the alternating cur- 
rent and it must, therefore, be operated on a direct-current supply, 
capable of producing about 1250 to 1300 volts. While its light output 
includes some more red than the other types of mercury lamps, it still 
is not entirely suited for color photography. It requires considerable 
auxiliary equipment in addition to the transformer in order to control 
the water flow. 

Fluorescent Lamps 

This source is not at all practical for high-speed photography be- 
cause of its relatively low brightness; in fact, it is not possible to 
obtain more than about 600 foot-candles on a surface even quite close 
to the lamps, irrespective of the number of lamps employed. 


In order to make this summary of light sources complete, flashtubes 
have been included in the electric-discharge types of lamps; however, 
their characteristics are quite completely covered by Edgerton, 1 and 
no further mention need be made of them here. 

Photoflash Lamps 

It may seem a little odd to mention the photoflash lamp which is 
essentially a flash source in a paper dealing with continuously burning 

40 FARNHAM March 

illuminants. However, one investigator 2 has taken advantage of 
the long flash duration of the FP-Type of flashlamp and combined 
the light output of a number of lamps, flashing in succession, to pro- 
duce a reasonably constant source of very high value. 

The General Electric Type 31 flashlamp has a flash duration of ap- 
proximately 55 milliseconds between J /2 peak values of the flash. It 
is possible, by a suitable motor-driven switch to initiate the flash of a 
second lamp in such a manner that the decay of the first lamp com- 
bines with the increase in light output of a second lamp to produce 
a continuous source ranging in value from 1V 4 to l*/2 million lumens. 
This process is continued with a third and a fourth lamp, and so on; 
thus 17 lamps will produce illumination lasting for one second. 

Equipment developed by Lester 2 mounts these 17 lamps on the 
periphery of a wheel which is driven by the same motor that operates 
the flashing contacts and which makes one revolution per second. 
Each lamp in succession passes through a slot at the back of a large 
deep-bowl reflector while going through its flashing cycle. 

This type of ilium inant possesses several unique advantages: 

An average of 1,400,000 lumens is emitted during this prolonged 
flash which, if obtained from lamps of the photoflood type, would re- 
quire 45 kilowatts of lamps. 

A couple of dry cells are all that are required to ignite the flashes, 
plus additional power to drive the small motor rotating the wheel on 
which the lamps are mounted. 

Thus, the use of photoflash lamps permits making high-speed 
pictures when the usual or sufficient power supply is not available. 

The heat received on the area being photographed is negligible, 
because of the short tune that the lamps are on . 

Two disadvantages might be mentioned. 

1. The lamps move some little distance through the reflector dur- 
ing the flash; thus, the direction of the light beam from the reflector 

2. The tune required to relamp after each shot. 


(1) Harold E. .Edgerton, "Electrical flash photography," /. Soc. Mot. Pict. 
Eng., this issue, Pt. II, pp. 8-24. 

(2) Henry M. Lester, "Continuous flash lighting An improved high-intensity 
light source for high-speed motion picture photography," /. Soc. Mot. Pict. 
Eng., vol. 45, pp. 358-369; November, 1945. 



QUESTION: When will the lamp be commercially available? 

MR. R. E. FARNHAM: It will be available very soon. 

QUESTION: What is the code number? 

MR. FARNHAM: 750R40. 

MR. W. T. WHELAN: Do you plan to bring this lamp out with a diffuser face 
like your 500-watt photoflood? Are you going to manufacture it with a diffuser 

MR. FARNHAM: The face of the lamp is clear. This has been done to pro- 
duce extreme concentration. The filament is of such a character that it will 
give a reasonable uniformity over the limited spot it is designed to cover. 

MR. WHELAN: We have an application where a very large area must be 
covered with fairly uniform diffused light and we use the 500-watt photoflood 
behind a secondary diffusing screen. 

I suggest that it might be useful to some users of the light sources to manufac- 
ture that particular type of lamp in the diffuser type. We use the one that is 
right above it on your chart. 

CHAIRMAN L. R. MARTIN: As Mr. Farnham explained, I think that this specific 
lamp was designed to specific requirements at the request of the Society's Com- 
mittee on High-Speed Photography. I feel that we all recognize that everything 
fits into the pattern, and I would not say Mr. Farnham would want not to 
recommend it for the type of service of which you speak. 

MR. FARNHAM: You are quite right. This lamp was designed for a specific 
purpose and you will find other lamps shown in that table that will meet the 
requirements for illumination on a larger area. 

Motion Picture Equipment for 
Very High-Speed Photography 



Summary With the increased use of high-speed photography to study 
various commonplace phenomena which take place at extremely high 
speeds, it has been found necessary to develop motion picture camera 
equipment capable of exposing film at increasingly greater rates. A brief 
description is given of various types of cameras capable of speeds up to a 
million .frames per second. 

MOTION PICTURE CAMERAS such as the Eastman high-speed or the 
Bell Laboratories Fastax, which use a rotating prism as optical 
compensator, possess so many advantages for studies at speeds up to 
about 12,000 frames per second that they are the almost universal 
choice for this range. However, a significant number of events of 
technical importance produce motion too fast* to record properly 
at 12,000 frames per second. For example, the simple spreading of a 
crack in a plate of glass occurs at a speed of about one mile per second. 
If a 10-inch square of window glass be struck in the center, the cracks 
will travel to the edge of the plate in V^.ooo second or in the interval 
between successive frames. Many commonplace electrical phenomena 
take place at even higher speeds. A piece of fine fuse wire subjected 
to a very heavy overload can expand at a speed of 10 miles per second! 
For the study of these and other very fast events, motion picture cam- 
eras operating at speeds of 50,000 to several million frames per second 
are required. This is the range covered in this paper. 

Methods of optical compensation other than the rotating prism 
may be used to permit continuous motion of the film, as in the rotating 
lens turret of the Merlin-Gerin camera and a camera very recently de- 
scribed by Partch and Wyckoff, 1 but these also suffer from the limita- 
tion in the speed at which the film may be moved. Consider a stand- 
ard 8-mm motion picture film, with 80 frames per foot. At 12,000 
frames per second such film must be moved through the camera at 150 
feet per second, a speed approaching the present limit for perforated 
film on sprockets. If greater speed is required some other method of 
handling the film must be provided. 



A significant gain in linear film speed may be obtained by attaching 
a single turn or a spiral of film to the outside of a rotating drum. For a 
given linear speed the centrifugal force varies as the inverse power of 
the radius of the drum, but even with large drums the problem of 
holding the film in place against the centrifugal force is quite serious. 
A marked mechanical improvement results from placing the film on 
the inside of a rotating drum, the centrifugal force holding the 
film in place. The optical arrangement with this system is not in 
general as convenient, but centrifugal force upon the film is no longer 
a problem, and the speed limitation is imposed only by the strength of 
material forming the drum. In this case no gain in linear speed re- 
sults from increasing the diameter of the drum. The mass of the 

Fig. 1 Scophony high-speed camera. 
D, rotating drum; F, film; M, multiple mirror; L, 
camera lens. 

drum, considered as a spinning ring, increases with increasing diame- 
ter so as just to offset the decrease in centrifugal force. Cameras 
handling film by this method provide the highest linear film speeds 
attainable to date. 2 Very recently a camera handling film in this 
manner has been reported by Scophony Limited. 3 This camera is 
illustrated diagrammatically in Fig. 1. A rotating multiple-face 
mirror M rigidly connected to the drum D carrying the film provides 
optical compensation for the motion of the film and yields an actual 
motion picture record. The Scophony camera operates at a film 
speed of more than 600 feet per second resulting in standard 35-mm 
motion picture frames at a speed of 10,000 per second. If 8-mm in- 
stead of 35-mm frames were used, with an appropriate alteration in 
the mirror, it would seem possible to obtain 50,000 frames per second 
by this principle. In effect this has been accomplished in the Suhara 1 




camera illustrated diagrammatically in Fig. 2. Here the optical com- 
pensation is provided by a multiple-face mirror M, gear-driven from 
the main shaft which, again, uses a shallow open drum D carrying 
within it a single turn of film. Speeds in excess of 100,000 frames per 
second have been reported from this camera, but the device is very 
large and cumbersome and has found little application. 

Quite recently Baird 4 has described a moving-film camera with opti- 
cal compensation provided by a multiple-faced mirror in which the 

Fig. 2 Suhara camera. 
Z), rotating drum; F, film; M, mul- 
tiple mirror; G, gear-drive to multiple 
mirror; L, camera lens. 

limitation of linear film speed is not so serious. By increasing the 
width of the film and providing a number of identical objectives side 
by side, Baird has in effect connected a number of moving-film mirror- 
compensation cameras together, side by side. Suppose that there are 
five objectives and thus five rows of pictures on the moving film. The 
multiple mirrors which provide the optical compensation, although 
mounted together rigidly on a single shaft, are stepped apart in angu- 
lar position by one fifth the angular separation between successive 
faces of the multiple mirror for any particular row. Thus objective 1 
and multiple mirror 1 produce frames number 1, 6, 11, 16, etc., of the 


final motion picture, while objective 2 and multiple mirror 2 produce 
frames number 2, 7, 12, 17, etc. In this way five times as many 
frames per second are obtainable for a given film speed and size of image. 

For nonluminous events which may be illuminated by a spark or 
repetitive flashing lamp, satisfactory high-speed motion pictures may 
be secured without any compensation for movement of the film. Ed- 
gerton and his associates have produced excellent photographs by 
this method, but again the number of frames per second is limited by 
the speed at which it is possible to move the film. The repetitive 
flashing lamp has another use which has not been so generally recog- 
nized. If linked with any of the optical compensation cameras so as 
to flash in synchronism with the camera mechanism, a very short 
exposure time is provided for any nonluminous portions of an event. 
It is common experience that a frame speed of only a few thousand 
per second is quite adequate to analyze many fast motions, yet the 
exposure time of individual frames may have to be reduced to V5o,ooo 
second or less to eliminate blur resulting from motion of the object 
during a single exposure. By the addition of such a lamp system 
cameras operating in the range of the Eastman high-speed or the 
Fastax can be applied successfully to events which , at first inspection, 
would seem to be beyond the speed range of these cameras. 

An obvious way to overcome the difficulty of fast film movement is 
to allow the motion picture film to remain stationary and to sweep the 
image along the film with a rotating-mirror system. Very high linear 
image speeds are possible by this method which has been applied ex- 
tensively in a number of streak cameras for the study of sparks, explo- 
sions, and other self-luminous events. To the authors' knowledge, 
this method has not been employed in any camera producing an actual 
motion picture record, but only in cameras which yield a streak image 
resulting from sweeping the image of the changing event along the 
film. However, an interesting variant of this procedure has been ap- 
plied to secure actual motion picture records. at very high speeds, a 
system in which the aperture rather than the image is caused to sweep 
by the rotation of a mirror system. By way of introducing this sytem 
the simple multiple-camera technique will be considered. 

As far back as 1877, when Muybridge made his first series of photo- 
graphs of a running horse, use has been made of a series of still cam- 
eras exposed in sequence as a means of photographing motion. In 
recent versions use is made of an electrooptical shutter provided by the 
Kerr effect in a liquid such as nitrobenzene contained in a cell mounted 


between crossed polarizers. The optical efficiency of such shutters is 
low but very high speeds are possible. The most recent work with 
Kerr cells appears to be that of Zarem 8 in which three cameras are 
tripped in sequence by an electrical-pulsing arrangement giving an 
equivalent exposure of about second. In addition to shutter 
inefficiency (poor light transmission), this system limits the number 
of motion picture frames of an event to the number of complete 
lenses and shutters provided. Part of this limitation has been re- 
moved in the Miller and in the Bowen high-speed cameras. 

In the Miller camera 6 the film and 90 identical photographic 
lenses are stationary, and the aperture of a primary objective is swept 

Fig. 3 Optical principle of the Miller and the Bowen 


M, rotating mirror; F, film; Z/i, primary camera lens; 
Z/2, field lens; Ls, secondary camera lenses. 

in succession across the 90 secondary objectives by a rotating-mirror 
system. In the Bowen camera 7 a somewhat similar arrangement is 
used to sweep an aperture across 76 separate lenses. Both the Miller 
and the Bowen cameras give speeds of about 400, 000 frames per second, 
and this does not appear to be the limit attainable. The fundamental 
principle of these cameras is illustrated in Fig. 3. An image of the 
event to be photographed is formed at the surface of .the rotating 
mirror M by a primary objective LI. By means of a field lens 
Z/2 an image of the aperture of the primary objective L\ is formed 
upon one of the stationary secondary objectives L 3 behind each 
of which is an area of photographic film. The only moving part 
is the mirror M , the rotation of which causes the image of the aperture 


of Z/i to sweep successively across the apertures of the secondar}' 
objectives L 3 . As the aperture of LI is imaged successively upon each 
of the lenses L 3 , that lens forms an image of the mirror (and thus of 
the event to be photographed) upon the film F. With the arrangement 
shown in Fig. 3 the number of frames in the final motion picture is 
equal to the number of secondary objectives L 3 . However, in the 
Miller camera by an arrangement of the multifaced mirror M it is 
possible to secure more frames than there are lenses L 3 in the secondary 
set. In spite of this the total number of frames in any run is quite 
limited, although sufficient to record a number of fast events. 

The required film or image speed in a moving-film camera can be 
much reduced if the conventional shape of the motion picture frame is 
altered to make the dimension along the direction of film movement 
very small. This has been done in a camera recently described by 
O'Brien and Milne 8 which combines very high speed with long runs of 
several thousand frames in a simple mechanical system. The change 
of shape from the conventional motion picture frame is accomplished 
by a stationary optical system called the image dissector. This opti- 
cal system, without moving parts, forms an image of the subject in 
the usual proportions of a motion picture frame, then cuts this image 
into a series of narrow strips, and redisposes these strips end to end to 
form a single long-slit image which is reimaged upon moving film with 
the long dimension across the direction of motion. All the informa- 
tion in the rectangular picture is contained in this narrow strip. It is 
only necessary that the negative film move the small width of this 
strip to record a new frame. By driving the film 400 feet per second 
on the inside of a shallow rotating drum, a camera speed of ten million 
frames per second is achieved. If a shutter opening and closing at this 
speed were provided, combined with image movement compensation 
on the film, no loss in resolving power would occur. In practice, how- 
ever, the camera is used without a shutter so that the picture is 
blurred by the width of the slit image on the moving film. In the 
present model this width is only about Vioo nun, so the loss of resolution 
is comparable to the resolving power of a fine-grain negative material. 

After processing, the negative is projection-printed back through an 
optical system similar to that which formed it, and a standard 16-mm 
motion picture frame thus reconstructed from Yioo mm of film travel. 
The printer is automatic and delivers a standard 16-mm motion pic- 
ture print of any desired length. 

In the first version of the camera reported the motion picture frame 


is cut into 15 or 30 strips, and the picture quality is very poor. In 
the second version of the camera under construction a much improved 
image quality is expected. The present camera permits continuous 
runs of 1600 motion picture frames for a 30-element image dissection, 
while permitting very long runs of more than 60,000 motion picture 
frames for a 15-element image. Because of the nature of the optical 
system the camera works at the high relative aperture of //2.0, an 
aperture which has not been approached in any of the other ultraspeed 
cameras reported to date. If the Type II version of this camera comes 
up to expectations it should prove a useful addition in this field. 


(1) N. T. Partch and C. W. Wyckoff, Navy Department Taylor Model Basin 
Report R-345, June, 1948. Descriptions of many types of high-speed motion pic- 
ture cameras are given in a report by John Waddell, "High-speed motion picture 
photography," Bell Telephone Laboratories, 1947. 

(2) Brian O'Brien, "A high-speed running film camera for photographic pho- 
tometry," Phys. Rev., vol. 50, pp. 400-401 ; August, 1936. 

(3) Scophony Ltd., "The Scophony high-speed camera," Phot. J. t vol. 86B, pp. 
42-46; March-April, 1946. 

(4) K. M. Baird, "High-speed camera," Can. J. Res., vol. 24A, pp. 41-45; July, 

(5) A. M. Zarem, Press release May 8, 1948, from Naval Ordnance Training 
Station, Inyokern, California; see also Electronics, vol. 21, p. 164; July, 1948. 

(6) C. D. Miller, National Advisory Committee for Aeronautics Technical Note 
140-S, June 10, 1947; see also J. Phot. Soc. Amer., vol. 14, p. 669; November, 1948. 

(7) Naval Ordnance Test Station Reprint 1033, Inyokern, California, "The 
Bowen 76-lens camera," by J. S. Stanton and M. D. Blatt. 

(8) Brian O'Brien and Gordon Milne, "Motion picture photography at ten 
million frames per second," J. Soc. Mot. Pict. Eng., vol. 52, pp. 30-41; January, 

Methods of Analyzing 
High-Speed Photographs 



Summary Qualitative, quantitative, and graphical methods of analyzing 
high-speed photographs are discussed and their peculiar requirements enu- 
merated. Apparatus adapted to the several methods are described in detail. 


THE METHODS of analyzing high-speed photographs may be con- 
sidered of three broad and somewhat overlapping types : quali- 
tative, quantitative, and illustrative or graphical. Each presents 
peculiar problems which can be simplified by suitable apparatus. 
In nearly all analyses orientation and time-scale references are de- 
sirable. In quantitative work they are, of course, essential. These 
references take many forms. They may be simply a linear scale, a 
single wire or two points of reference, and a time clock in the field. 
In more elaborate form, an over-all grid or checkerboard background 
and accurately timed spark discharge or argon glow lamps to produce 
pips on the edge of the film, provide orientation and timing references. 


To see the changing relationship of elements of a high-speed 
mechanism, the relative movement of fragments of a disintegrating 
subject, or to determine whether a particular phenomenon does or 
does not occur is readily accomplished by projecting a motion picture 
film taken at high speed with a conventional motion picture projector. 

Such a projector should be capable of forward and reverse motion 
and still-picture projection. Remote controls are advantageous and 
a frame or footage counter is often very useful. When the exposure 
rate is known and the projection rate can be suitably adjusted and 
controlled, the duration of a cycle or a transient phenomenon can 
sometimes be measured with a stop watch. 

Typical of projectors for qualitative analyses is the Kodascope six- 
teen-20, Fig. 1, which incorporates a push-button control panel, verti- 
cal-tandem pulldown claw, and forward and reverse and still-picture 





projection. Standard equipment includes a two-inch //1. 6 coated 
lens, a 750-watt lamp, and a carrying case. A 1000-watt lamp is 
available for this model. The carrying case also serves as a projection 

Another 16-mm projector is the Bell and Ho well Model 173 AD, 
Fig. 2. Designed especially for time and motion study it provides a 
two-perforation pulldown claw, forward and reverse and still-picture 
projection, a single-frame hand crank, and a frame counter. The 
electric governor-controlled motor provides a speed range from about 
13 to 20 frames per second. Standard equipment includes a 2-inch 

//1. 6 coated lens, a 750-watt 
lamp, and a carrying case. A 
1000-watt lamp may also be used 
in this model. 

Shown in Fig. 3 is a 16-mm 
Keystone projector Model A-82, 
as modified by Professor David 
B. Porter of New York Univer- 
sity to adapt it for industrial time 
and motion study. To reduce 
the inertia of the shutter and 
thereby minimize coasting and to 
reduce the flicker rate, which in 
all projectors becomes very ob- 
jectionable at two or three frames 
per second, the standard shutter 
is replaced by a single blade of 
0.010-inch Duralumin. The 
standard motor switch is re- 
placed by one of the more easily operated key types. A toggle switch 
is added for the projection lamp and one for a pilot. The pilot lamp 
also serves to illuminate work papers. The motor rheostat and all 
controls are mounted in a small box and provided with a 6-foot cable 
for remote operation. A further useful addition is a frame counter. 
With these modifications the projector can be operated at very low 
rates and the film can be advanced frame by frame. 

Similar Kodascopes, Filmo, and Keystone projectors are available 
for 8-mm film. 

Another useful device is the "editing viewer." Mounted between 
rewinds and driven by the film, it projects a small motion picture on a 


1 Eastman 16-mm Kodascope 




translucent screen. The use of a rotating prism, as in high-speed 
cameras, permits continuous motion of the film at low or high speeds. 
Usually incorporated in such a device and of particular value, is a 
means of marking individual frames for later detailed examination or 
the making of single-frame enlargements. 

Fig. 2 Bell and Howell 16-mm projector Model 173 AD. 

A desirable adjunct to the viewer is a footage or frame counter. 
A counter facilitates indexing and finding particular sections of the 
film and by reference to a velocity-footage curve the rate at any 
point in the film can be determined. One source of such counters 
is the Neumade Products ^Corporation. 

The Bell and Howell filmotion viewer is shown in Fig. 4 with their 




Model 136 splicer and rewind assembly. The viewer employs a 30- 
watt line-voltage lamp and projects an image approximately 2 1 /* by 3 
inches. Controls are provided for framing, focus, and a marking 
device which produces a tiny slit in the edge of the film when it is 
desired to mark an individual frame. The rewinds shown accommo- 
date 2000-foot reels. They have two-speed crank drives and are 
equipped with brakes. 

Fig. 3 Modified Keystone 16-mm projector Model A-82. 

Fig. 5 illustrates the Cine" Kodak editing viewer as part of the 
master editing outfit which also incorporates a splicer and 1600-foot 
brake-equipped rewinds. This viewer projects an image approxi- 
mately I 1 / 'a by I 1 / 2 inches. Focus is fixed and controls are provided 
for framing and an edge-notching device for marking individual 

Similar editing viewers are also available for 8-mm film. 





Having, by previous inspection, determined that certain sequences 
or individual frames will yield additional data by quantitative study, 
other apparatus will be found especially useful. 

All of the equipment hereinafter described is drawn from the 
seemingly unrelated field of microfilming. Film readers, as they are 

Fig. 4 Bell and Howell 16-mm filmotion viewer. 

Fig. 5 Eastman 16-mm Cine Kodak editing viewer. 

generally called, are, however, specifically designed to facilitate the 
rapid handling and critical examination of discrete images on 16-mm 
and 35-mm film. The very nature of their use in reading microfilm 
requires the production of screen images of excellent quality. High 
resolution, ample and uniform illumination, simplicity, and ease of 
operation are requisite to long periods of use without physical and 
visual fatigue. 




By way of comparison with a probably more familiar subject, it 
can be stated that in current practice higher values of resolution are 
obtained in microfilming than in sound recording. A film commonly 
used for microfilming is rated at just double the resolving power of 
fine-grain release positive, and microfilm negatives commonly yield 
resolution higher in the same degree. All of these comments on micro- 
filming are, of course, merely to qualify and recommend such film 
readers for the analysis of high-speed photographs. 


Fig. 6 Recordak 16- and 35-mm film 
reader Model C. 

Fig. 7 Recordak 16-mm film reader 
Model PD. 

One of the most widely used film readers of that type is the Re- 
cordak film reader Model C, Fig. 6, which accommodates both 16- 
and 35-mm film. In this device images are projected on a self-con- 
tained 18- by 18-inch screen of special plastic composition which 
minimizes scintillation and hot spot so objectionable in ordinary 
ground-glass screens. Magnifications from 12 to 23 diameters are 
afforded merely by moving the screen outward on its hinged support. 
The image is automatically focused throughout the range. Variable 
magnification and automatic focusing facilitate adjustment of the 
image size to match a preselected scale or re*seau. 




Since at high magnification the 18-inch screen does not contain the 
full width of 35-mm film, a scanning mechanism is provided to move 
the film with respect to the projection lens and thus to bring any por- 
tion of the frame into view. The image can also be rotated on the 
screen through 360 degrees. By these means reference points in the 
image can be brought into coincidence with similar points on an over- 
laid screen pattern. 

Illumination is obtained from a prefocused 200-watt lamp and the 
film is protected by a heat filter and by being clamped between glass 
flats. The flats are opened automatically when the film is progressed 
in either direction by means of the low- or high-speed hand cranks. 

Fig. 8 Recordak 16- and 35-mm portable projector 
Model A. 

An accessory shelf is available to provide a work surface in front o 
the screen. 

Another Recordak film reader, the Model PD shown in Fig. 7, is 
designed to project 16-mm film at a magnification of 23 diameters. 
Its self-contained translucent screen measures 14 by 15V2 inches and 
is made of the same special material as the screen in the Model C. 
Other similar features are the 200-watt lamp, the heat filter, film- 
clamping flats, and image rotation. In addition, the Model PD can 
be used as an enlarger. A paper holder is incorporated in the unit 
and it can be readily adjusted for magnifications from 8 to 18 

In Fig. 8 is illustrated the Recordak portable projector Model A 
which accommodates either 16- or 35-mm film. Magnifications from 

56 NIVISON March 

10 diameters upward can be obtained, with 30 diameters possible at 
6 x /2 feet from the screen. Heat-filter and film-clamping flats are in- 
corporated. The flats are manually operated. A projector of this 
type readily lends itself to the arrangment described by Thomas and 
Coles 1 wherein a mirror is placed in the beam of the projector to reflect 
the image back to a translucent screen near the projector so that the 
analyst can operate the projector and measure the screen image from 
the same position. 

Fig. 9 Threading the Recordak transcription film reader. 

Being readied for production is a new Recordak transcription film 
reader which incorporates many novel features. Designed to be used 
in making transcriptions by hand, typewriter, card punch, or posting 
machine its development was directed toward ease of operation, ideal 
placement of the screen, flexibility, and convenience of use. 

The base resembles a desk-high filing cabinet with the screen 
mounted in a rotatable turret and the projection mechanism contained 
hi the top drawer. Knobs on the face of the drawer control image ro- 
tation, focus, and lateral scanning of the film. A remote control for 
the motor-driven film transport is also included. Placed adjacent to 
a standard desk, as shown in Fig. 9, reel spindles are accessible and 




threading extremely simple and convenient. Magnification can be 
varied between 24 and 35 diameters merely by moving the drawer in 
or out. 

Fig. 10 shows how the screen turret can be rotated to overhang the 
work surface and that the screen and work papers are equidistant from 
the analyst's eyes. Thus by minimizing accommodation, visual fa- 
tigue is reduced. The remote control provides three rates of film 
transport, both forward and reverse, and the two lower rates are 

Fig. 10 Transcribing data with the Recordak transcription 
film reader. 

adjustable. Frame-by-frame advance also is possible. Of significance 
is a new optical device which replaces the commonly used film-clamp- 
ing flats. This device holds the film in critical focus whether station- 
ary or in motion and without damage to the film after many thousand 


Some analyses can be accomplished more readily with single-frame 
enlargements. Enlargements too are often desired for publication, 
demonstration, or record purposes. While conventional erilargers will 
sometimes suffice, here again microfilming equipment is eminently 
adapted to the task. 




Many industrial organizations 
maintain photoreproduction de- 
partments and employ Photostat, 
Rectigraph, or similar photocopy- 
ing machines. Microfilm en- 
largers have been designed for at- 
tachment to such machines to take 
advantage of the paper-handling 
and processing facilities which 
those machines incorporate. In 
the larger, so-called "continuous," 
machines completely processed 
and dried prints are delivered at 
rates of more than one per minute. 

Fig. 11 shows a Photostat microfilm enlarger as mounted on the 
front of a typical Photostat machine. In this position it replaces the 
regular lens-and-prism assembly and projects an image either to a 
viewing screen or back onto the sensitized paper. The device com- 
prises a projection-lamp unit, condensing lens, projection lens, prism, 

Fig. 11 Photostat 16- and 35-mm 
microfilm enlarger. 

Fig. 12 Haloid 16- and 35-mm microfilm enlarger. 




focusing scales, and a manually operated film-winding mechanism. 
This enlarger is offered with one or two projection lenses and a magni- 
fication range from 5 to 30 diameters depending on the lens equip- 
ment and the model of the machine to which it is attached. 

Fig. 13 Recordak 1&- and 35-mm en- 
larger Model A. 

Another form of microfilm enlarger is designed for use on the Recti- 
graph and other photocopying machines. It can be used independ- 
ently of such machines if a suitable paper-handling mechanism is pro- 
vided. Equipped with a motor-driven film transport, it incorporates 
a photoelectric control which advances the film, frame by frame, 



regardless of the length of the frame (microfilm frames often vary 
in size). The control operates on the difference in density of the ex- 
posed frame and unexposed frame line. 

With this enlarger mounted on a Haloid Foto-Flo machine, Fig. 12, 
and equipped with an interlocking control, enlargements are made 
automatically at rates from 4 to 10 per minute depending on the size 
of the enlargements. Magnification range is from 8 to 26 diameters. 
The enlarger can be swung out of the way to permit normal photo- 
copying operations. This enlarger, designed by the Microtronics 
Corporation for Haloid, is sold exclusively by the Haloid Company. 

The Recordak enlarger Model 
A shown in Fig. 13 provides mag- 
nifications from 4 to 30 diameters 
with either 16- or 35-mm film. 
Its 63-mm Ektar lens has an ex- 
tremely flat field and high resolv- 
ing power. Critical focus is as- 
sured by clamping the film be- 
tween glass flats and by correlat- 
ing the lens scale and measuring 
tape which are both calibrated in 
diameters of magnification. 

The simple and inexpensive 
device shown in Fig. 14 is another 
form of single-frame enlarger. It 
is no longer manufactured but 

Fig. 14-Jiffy Kodak 16-mm enlarger. sti11 can be f ound in camera shops, 

and is extremely useful . A modi- 
fied, fixed-focus, Jiffy Kodak with a 16- or 35-mm film gate provides 
a convenient means of making enlarged negatives on 2 1 /*- X 3 1 /*- 
inch roll film. 

Obviously, this list is incomplete. Not all of the apparatus which 
might be used for the analysis of high-speed photographs have been 
mentioned but the author feels from his experience and that of other 
workers in the field the apparatus herein described can be commended 
to use in analysis work. 


(1) P. M. Thomas and C. H. Coles, "Specialized photography applied to 
engineering in the Army Air Forces," J. Soc. Mot. Pict. Eng. t vol. 46, pp. 220- 
231; March, 1946. 

New Developments in 
X-Ray Motion Pictures 



Summary Equipment using exposure times as short as 10 microseconds 
permits the radiographing of very rapidly moving objects and the use of con- 
tinuously moving film without blur in a specially constructed camera without 
a shutter. The system may also be used in some cases to produce an image 
on a fluorescent screen which may be photographed with a General Radio 
oscilloscope camera or a synchronized motion picture camera. 


WHEN HIGH-VELOCITY ELECTRONS impinge upon matter, X rays 
are produced. The higher the atomic number of the element 
struck, the higher is the efficiency of the X-ray production; also, the 
higher the velocity of the electrons, the higher the efficiency and also 
the shorter the wavelength of the resulting X rays. These shorter- 
wavelength X rays are more penetrating than those generated with 

Fig. 1 "Micronex" surge generator and field-emission 
X-ray tube making single radiographs with exposure times 
of Vw to 1 microsecond. 



the lower-velocity electrons. Thus, with X rays it is possible to con- 
trol both the quantity and quality of the rays produced merely by 
regulating the voltage and current through the high-vacuum X-ray 

At reasonable voltages, many times as much energy is required to 
produce sufficient X rays for a picture as would be needed with visible 
light, and since this energy must be supplied at high voltage, the 
equipment is, of necessity, larger and more cumbersome. X rays 

Fig. 2 Radiograph of lead dust passing through a vacuum 
cleaner. Exposure time one millionth of a second. 

cannot be refracted in the usual sense of the word, so that all X-ray 
pictures must be shadowgraphs, with the size of the picture equal to 
or larger than that of the object. If there is a sufficient quantity of 
X rays available, it is sometimes possible to cast an image on a fluores- 
cent screen and then photograph this screen with a high-speed camera. 
Some definition is lost in this process and depending somewhat on 
the quality of the X rays, their intensity must be 20 to 100 times that 
which is required to make the radiograph directly on X-ray film with 
the help of the usual intensifying screens. 


Since X-ray pictures are essentially shadowgraphs of the internal 
structure of objects, the definition is dependent upon the size of the 
X-ray source, which is the area of the target on which the electrons 
impinge. The ability of this focal-spot area to withstand the energy 
of the electron bombardment is a determining factor in controlling 
the intensity of the X rays which it is possible to produce. At any 


striking a ball. Note the compressed 
core. Exposure time one millionth of a 

given voltage, the quantity of X rays is proportional to the current 
through the X-ray tube so that the exposure time required for a 
picture will be inversely proportional to this current. 

Recent developments have produced a new type of electron emitter 
called a field-emission arc, which enables currents 100,000 times larger 
than was possible with the hot-cathode type of X-ray tube. This has 
permitted X-ray pictures to be made in the extremely short time of 1 
microsecond and very recent improvements have reduced this time to 
Vio of a microsecond. This new type of equipment, shown in Fig. 1, 




has been widely used for studying the action of very rapidly moving 
enclosed objects and under conditions where it is not possible to use 
the techniques of high-speed photography either because of the light 
produced or of the inability to protect the photographic equipment 

Fig. 4 Radiograph of a bursting model bomb showing 
distribution pattern of fragments. Radiograph made by 
the Ballistics Laboratory at the Aberdeen Proving Center 
in 1 microsecond. 

from flying fragments of bombs or bursting shells. With X rays, it is 
feasible to use armor plate to shield the X-ray tube as well as the 
photographic film. Figs. 2, 3, and 4 are typical applications of this 



As in the case of photography, when it was found possible to take a 
single picture with a very short exposure time, the question immedi- 
ately arose as to what happened just before or just after the picture 
was taken. In other words, what would X-ray motion pictures show? 
Under contract with the Navy, the development of X-ray motion 
picture equipment was initiated, intended primarily for the study of 
burning rocket propellants. Requirements to be met were 100 
frames per second, each exposure to be no greater than 10 micro- 

In order to insure the success of this project the development of two 
types of equipment was undertaken. The first followed more conven- 


Fig. 5 Simplified choke-charging pulse-transformer 
circuit used to energize the X-ray tube. This circuit is 
similar to a radar-type line-modulator circuit with the 
pulse-forming network replaced by capacitor C 2 . 

tional circuit procedures in which the high voltage was obtained 
through a choke-doubling circuit which charged the final capacitor to 
150 kilovolts. This was then discharged through the X-ray tube by 
means of a special triggering tube, the sequence being repeated 100 
times a second. The second method, which proved to be the more 
successful, used the radar-pulsing principle in which an especially de- 
signed pulse transformer capable of stepping up the 20-kilovolt supply 
to 150 kilovolts was the heart of the system. This arrangement 
enabled the energy to be stored at relatively low voltages in capacitors 
and then to be discharged through the primary of the pulse trans- 
former by means of a high-powered hydrogen thyratron similar to 




that used in radar techniques. Figs. 5 to 8 show the various compo- 
nents and assemblies of this equipment. 

For certain purposes where the penetration demands are not great 
and where it is possible to work fairly close to the X-ray tube, it has 
been found feasible to utilize the technique, previously described, of 
photographing the image on the fluorescent screen. For many 
applications, including that of the rocket studies, there is not sufficient 

Fig. 6 View of complete equipment; direct-recording camera on the left, 
oil-filled head containing the X-ray tube and pulse transformer in center, and 
power and control unit at the right. 

X-ray intensity to employ this technique satisfactorily and it was 
necessary to develop a special X-ray camera for this purpose. Fig. 9 
shows a photograph of the camera with the cover removed. An X-ray 
film 9*/2 inches wide and 25 feet long with 25 feet of leader and trailer 
is wound on the upper spool. During the exposure, lasting 1 second, 
in which 100 pictures are taken, this film is wound on the lower spool, 
passing around the large rotating drum covered with an X-ray intensi- 
fying screen which greatly augments the intensity of the image 


produced. Since the exposure time for each picture is only 10 micro- 
seconds, no shutter is required and there is no blur due to the motion 
of the film. Fig. 10 shows a single frame taken of a simulated rocket 
using this technique. 

Fig. 7 Rear view of power unit. 
Power supply and storage capacitors 
below, rectifiers in center section, and 
triggering thyratron in the upper 

A V4-horsepower motor is used to drive the film take-up spool and 
the film travels at 25 feet per second or roughly 17 miles per hour. 
At this speed precautions must be taken to maintain the proper 




tension so that the film remains in good contact with the screen and 
completes its run without damage. 

The developed film strip, 25 feet long by 9 J /2 inches wide, contains 
100 images. These may be examined individually, in detail, and the 

Fig. 8 View of power unit with door open showing sequencing 
and control components. 

velocity of the observed phenomena measured with the help of a 1000- 
cycle-per-second time scale imprinted on the edge of the film. If a 
visual impression of these phenomena is desired, the radiographs may 
be copied, frame by frame, on 35-mm or 16-mm motion picture film 




and projected in the usual manner at some suitable speed. Such a 
motion picture showing a Thermit reaction has been made which 
demonstrates the possibilities of this technique. Fig. 11 shows re- 
productions of several frames of this motion picture. 

Fig. 9 Direct-recording cam- 
era removed from its housing. 

Fig. 10 Contact print from section of X-ray film showing one frame of a 
simulated rocket. Timing markers along the edge are recorded at a rate of 
1000 per second. 


In ordinary industrial motion picture work each new study requires 
a somewhat different technique, and so it is with X rays; such equip- 
ment as this can deal with only a very minute portion of the problems 



requiring X-ra> motion pictures for their solution. It does, however, 
point the way and indicates that if the demand is great enough, X-ray 
motion pictures can be developed which will meet the requirements 
necessary for the solution of the industrial and ballistic problems 
which now cannot be completely solved by any other means. 

Fig. 11 Nine frames selected from a 200-frame series showing the reaction 
of Thermit (a mixture of iron oxide and aluminum) in a crucible. The molten 
iron is seen to form and run out the bottom of the crucible, below which it burns 
through a steel plate. The frame numbers indicate their position in the original 
series which was recorded at one hundred frames per second. 

Some other fields in which the X-ray motion picture is expected to 
furnish important information include metal flow in arc welding, 
certain types of chemical reactions, motions of 'solid and liquid ma- 
terials in agitators, closed conveying systems, gears and valve 
mechanisms, and the like. 

High-Speed and Time-Lapse 
Photography in Industry 
and Research 



Summary High-speed and time-lapse photography are related to each 
other by the common factor of time mobility in photography. Similar prob- 
lems exist in both types of photography, the most common of which concern 
lighting, timing, and interpretation. This paper discusses a number of cases 
in which these problems were attacked and solved by methods ordinarily con- 
sidered unorthodox. A short discussion is included on the application of 
some of these specialized techniques to conventional cinematography. 

IT APPEARS NECESSARY at the beginning of any discussion on indus- 
trial photography to distinguish between purely illustrative pho- 
tography (that intended to portray certain processes or products), and 
photography as a medium of research and investigation. Frequently 
films or photographs made for investigational purposes have turned 
out to be fascinating subjects in themselves, wholly aside from their 
content of scientific data. Conversely, the occasion often arises for the 
use of high-speed photography as an interpretive medium, to make 
abstruse processes or principles intelligible to the layman. It is in 
these cases where the ingenuity of the industrial photographer is 
taxed to the utmost, and the present paper concerns itself mainly with 
these specialized problems. 

Since we are concerned more with some of the unusual aspects of 
photography than with routine methods, it is well to point out here 
not only that unorthodox methods must frequently be used, but that 
these unorthodoxies often lead to improved methods of accomplishing 
certain photographic results in fields far removed from the original 
effort. In our work, however, unorthodoxy does not imply the use of 
highly specialized equipment in the main, though we have had occa- 
sion to devise certain units which are not generally available. More 
often, it involves the application of conventional apparatus in a some- 
what unconventional way. 

The scope of the industrial and research photographer's interests 
cannot be confined to any one type of work, high-speed included ; even 


72 LESTER March 

less to any one range of speeds. The fundamental utility of the cine 
ramera in industry and research lies in the independence of camera and 
projector speeds. The ability to vary the rate of operation of one 
while maintaining constant the speed of the other gives a mobility to 
time, and affords the camera a function much greater than that of a 
mere recording medium. High-speed photography is one facet of this 
function ; time-lapse photography is another. In a way, they may be 
thought of as opposite ends of a single spectrum. 

When the camera operates at the same speed as the projector, the 
result is nothing more than a record a mechanical notebook, so to 
speak. This is valuable, of course; a given action may be studied 
again and again, until it has delivered its full quota of information. 

Obviously, there exists the possibility of varying the speed of cam- 
era or projector to effect an extension or telescoping of the time ele- 
ment in the film. The possibility of projector-speed variation is pres- 
ent but highly limited speeds above normal by wear and tear on the 
film itself speeds less than normal by discontinuity and flicker in the 
projected image. 

But with the projector considered as the constant-speed reference, 
there is no necessity to limit the speed range in which the camera may 
be operated. Films taken at rates above normal and projected at nor- 
mal speed are necessarily seen as slowed down; we prefer the term 
"time-extended" films. This would appear to be self-evident, yet 
much confusion exists among laymen when high-speed photography is 
discussed. Frequently we have been called in to discuss a proposed 
high-speed study, only to find that what was really desired was a 
speeded-up film of some very slow action; our term for this is "time- 
telescoped" film. 

This latter form of time translation is a valuable research tool ; it 
must be realized that many processes, particularly in chemistry and 
biology occur too slowly for study at their natural rate. For this 
reason, as industrial and scientific photographers, the scope of our 
work includes normal-speed, high-speed, and time-lapse photography. 
The. value of this wide-range interest lies, as will be seen, in the fre- 
quent adaptation of techniques outside their usual fields, often a 
richly rewarding experience. 

This may be summarized, then, by the simple statement of purpose : 
we are concerned with the mobility of time as it can be effected by the 
camera. Our problems involve means for mechanically translating 
the actual speed and duration of motion in time into terms that are 
within the range of our limited powers of visual and mental perception. 



Strictly speaking, high-speed photography may be defined as any 
motion picture photographed at rates higher than that at which it will 
be viewed later. In our case, the term signifies a definite range of 
taking speeds. The common "slow-motion" camera operates at rates 
4 to 8 times normal, that is from 64 to 128 frames per second. At the 
other extreme, there are image-dissecting cameras with speeds rang- 
ing around the rate of 1,000,000 pictures per second. 

Our work mainly lies in a middle ground between the two extremes. 
Industrially, at least, speeds above 300 frames per second yield the 
most information, and we have found little to be gained by speeds 
above 3000 pictures per second. Most of our work has been done in 
even a more narrow range, using the Eastman high-speed camera, 
Type III, with a usable speed range of 1000 to 3000 frames per second. 
These speeds correspond to time magnifications of 65 to 200 times. 
We have made no important modifications in the camera itself, which 
has been fully described in the literature. 1 

The major problem in the use of this camera is strictly of a photo- 
graphic nature. Operating at its peak speed of 3000 frames per sec- 
ond, and having an effective shutter angle of 72 degrees, the camera 
yields an exposure on each picture of only Vw,ooo second, which poses a 
serious lighting problem. 

Familiar light sources, such as reflector-photoflood (RFL-2) and 
reflector-spot lamps (RSP-2), are conventionally used with this cam- 
era. Their effective area coverage is measured in square inches, with 
even the fastest 16-mm film and large-aperture lenses. Industrial 
problems involving the photography of larger fields require kilowatts 
of light, frequently demanding special wiring and auxiliary transform- 
ers. Even when such facilities were available, the resulting films were 
of dubious value. 

The validity of such results is doubtful because great quantities of 
heat accompany the enormous power dissipation of these large lamp 
concentrations. With industrial subjects, we were never sure that 
observed performance on the screen was actually inherent in the sub- 
ject; there was always the suspicion that some aspects of the action 
may have been caused by exposure to heat. Biological subjects were 
frequently withered or killed before they could be photographed, and 
even if they were not, a definite reaction to heat was always evident. 

Our eventual solution of the light-versus-heat problem was the 
adoption of a high-output, short-duration illuminant, the familiar 

74 LESTER March 

still photographer's flashlamp. We found that the standard General 
Electric focal-plane flashlamp 31 has a light output lasting nearly 70 
milliseconds, with a nearly level plateau of almost 56 milliseconds. 
At camera speeds of 3000 frames per second, this is enough to expose 
nearly 4 feet of film. Then 17 such flashlamps, fired in succession 
with suitable overlap, will provide one full second of continuous light, 
substantially uniform in level. 

Fig. 1 The author photographing a drone fly with high-speed 
camera and continuous flash-lighting units. (Photo by David B. 
Eisendrath, Jr.) 

A pair of continuous flash-lighting units (described elsewhere 2 ) pro- 
duce about 3,000,000 lumens for a period of one second. This is sev- 
eral times the brilliance of normal July sunlight. When the built-in 
tinier on the camera is set to start the flashing after 35 to 40 feet of film 
have run through, the lamps in both'units will continue flashing to ex- 
pose the remaining 60 or 65 feet of the film as it attains the speed of 
3000 frames per second. 

As for exposure level, two units, each 24 inches from the subject, 
cover an area more than 2 feet square, with sufficient light to expose 
fast black-and-white film at //ll to //16, with the camera at peak 
speed of 3000 frames per second. Color film can be exposed quite ade- 
quately under the same conditions at //2.7 to//4.5. The heat output 


is trifling; what heat is radiated in the brief flashing time is blown 
away by the whirling lamps. Equally important is the factor of 
power consumption ; the 2 units require only a 6- volt dry-cell battery 
for the flashing circuit, and less than 100 watts of 115- volt alternating 
current for the operation of both motors. 

Even when only a small field is to be photographed, as in the drone- 
fly photograph shown in Fig. 1, the tremendous light output of these 
units is of inestimable value. The live subject was little affected by 
the insignificant heat output of the units, and the depth of field and 
sharpness obtainable with long-focus lenses stopped down to //12.5 

Fig. 2 Frame enlargement of drone-fly film made with 
setup shown in Fig. 1. Note unusual depth of field and 
definition possible when adequate illumination permits 
stopping down to //12.5. 

provided a new experience in the photographic quality and clarity of 
high-speed films. Fig. 2 is an actual enlargement of a frame of film 
made in this way note the unusual sharpness and detail, particularly 
of the hairs covering the fly's body. This was photographed at 3000 
pictures per second, on 16-mm film. 

The value of such light sources for strictly scientific work cannot be 
overestimated; they are, however, equally important in another way. 
We have been able to work with as many as four such units; with the 
ample reserve of light available we were no longer limited to concen- 
trating their beams frontally for the mere sake of adequate exposure. 
Lighting was arranged as for any normal photograph, for optimum 

76 LESTER March 

modeling and photographic quality. The result may be said to have 
introduced a new realization of clarity into high-speed photography. 
Fig. 3 shows one such setup in which three units are visible, the 
fourth being behind the press for back light. 


Making a high-speed photographic study of a subject is only one 
part of the problem. The finished film must be interpreted by the 


Fig. 3 Group of continuous flash-lighting units used in 
photographing impact extrusion process. 

user, usually by projection, and frequently as well by study of single- 
frame enlargements, or frame-by-frame examination of the film itself. 

Engineering studies are usually turned over to the client's staff for 
whatever use they may wish to make of them; our aid in this case is 
limited to providing a time base or reference where it may be required. 
This may take one of two forms : either a millisecond clock is included 
in the scene being photographed, or an argon timer built into the cam- 
era provides visible markers at intervals of Vi2o second along the edge 
of the film. 

Films designed for nontechnical audiences pose a more difficult 


case. To an engineer, a slowed-down film of a machine is easily dis- 
tinguishable from a normal film of the same machine running slowly. 
The difference, generally, lies in the presence of certain second-order 
effects, vibration, for example. Frequently the study of these second- 
order effects is the very objective of the high-speed investigation. 

The layman, on the other hand, has little conception of such subtle 
differences, and his imagination boggs down at millisecond clocks. It 
is necessary in films aimed at the lay public to provide a frame of ref- 
erence in more familiar terms. We have found it possible to accom- 
plish this by careful choice of comparison objects, usually taken from 
nature. Thus, in a recent high-speed film made for a major industry, 
one in which we investigated the performance of the 750-ton high- 
speed power press alreadj' shown, we went far afield for comparison. 
Our reference object in this case was the tiny hummingbird whose 
wings beat 70 times per second. 

Dramatic as the comparison was, it alone was not completely self- 
explanatory. For complete comprehension (and for those who had 
never seen either the hummingbird or the high-speed power press) we 
included several feet of normal speed (16 frames per second) photog- 
raphy of each. Thus, both the subject and the reference object were 
presented, each at its normal operating speed, and slowed down 200 
times by the high-speed camera. Only in this way could we be sure 
that the audience would understand thoroughly not only the demon- 
strated process but also the method by which it was presented. 


While there may seem to be little apparent relation between high- 
speed photography and time-lapse studies, we have found it impor- 
tant to couple both techniques; first, because they are the logical ex- 
tensions of the principle of time mobility and second, because some 
techniques applied in one field stem directly from the other. 

For if rapid action may be slowed down extended in time perhaps 
is a better concept for more convenient study by the simple expe- 
dient of photographing it at rates higher than normal, then obviously 
the reverse procedure is equally relevant. Many industrial processes 
are extremely slow; either their motion is practically invisible, or a 
cycle extends over so long a period as to make study tedious. The 
same may be said for chemical processes and many natural phenomena. 

For purposes of definition, rather than limitation, we may charac- 
terize time-lapse photography as covering anything photographed at 

78 LESTER March 

rates lower than projection speed. In our work, this may mean at 
any rate from 8 exposures per second to a single exposure in 24 hours, 
though this by no means exhausts the possible range. Again, there is 
an optimum covering the majority of cases; our rule of thumb is 
based on the 100-foot roll of 16-mm film, having a projection time of 
slightly less than 4 minutes at 16 frames per second. Such a roll con- 
tains 4000 separate frames of film, and generally it is found conveni- 
ent to divide the cycle to be studied into 4000 parts and use one such 
part as our time interval. This, however, is not a rigid rule; for more 
rapid tempo, particularly when the film is intended for lay audiences, 
we frequently use shorter rolls or units, perhaps 1000 or even 500 
frames for a complete cycle. 

Illumination, so far as quantity is concerned, is not a problem in this 
field; heat, however, always is. A cool light source is preferred, and 
we have found that the electronic condenser discharge flashtube, or 
so-called "strobe-light," meets every requirement of an ideal illumi- 
nant for time-lapse photography. Originally intended as a light 
source for high-speed photography with the shutterless type of cam- 
era, and later monopolized as a dependable light source for still pho- 
tography, its remarkable uniformity, low heat radiation, and high ef- 
ficiency recommend it for use as a light source for intermittent motion 
picture studies. 3 

The camera used for this work is a conventional Cin6-Kodak-Spe- 
cial, operated by a simple intermittent motor and timed by the Ko- 
dak interval timer. The only auxiliary equipment required is a sim- 
ple commutator which triggers the flashtube at the open point of the 
shutter cycle. The commutator is connected to the single-frame-per- 
turn shaft of the camera. 

Photographically this is an elegant solution for a number of rea- 
sons. First, the light is not only heatless, but remarkably uniform 
from one exposure to the next. Second, the exceedingly short flash 
duration eliminates the possibility of flicker or blur caused by uneven 
camera speed or backlash in the shutter gears. Third, the exposure 
required is amazingly predictable; a simple mathematical formula 4 
suffices, when the constants of the power pack are known, to deter- 
mine the required exposure with a high degree of accuracy. 

The problem of interpretation exists in the case of compressed-time 
films also. Unlike high-speed films, however, time-lapse photography 
is seldom studied frame by frame. The problem in this type of work is 
not to slow down motion so the subject can .be studied, but to 




compress the time so that movement can be perceived. For this reason 
the time base, when necessary, should be included in the film; an or- 
dinary clock or watch usually serves the purpose. 

Again, for nontechnical audiences, a frame of reference is required, 
and again we go to nature for our comparison. In one instance the 
subject chosen as a comparison base for a chemical process was a talis- 
man rose, photographed at 2-minute intervals for 5 days and 5 nights. 
Protected from light and air currents, the rose developed undisturbed 

Fig. 4 Rose, shielded from light and air currents, is photographed 
at 2-minute intervals for 120 hours by light of bare FT-214 flash- 

by any external influences (Fig. 4). The resulting film showed in 
about 3 minutes on the screen a smooth, steady progression from bud 
to bloom and finally a withered specter, unaccompanied by the usual 
gyrations of a flower subjected to alternate periods of daylight and 

Filming in this case was on daylight-type Kodachrome film, which is 
spectrally well matched to the light of the electronic flashtube. Though 
this film is extremely sensitive to variations in illumination level, 
no visible flicker resulted from the use of an intermittent light source. 



As already mentioned, the unorthodox use of conventional equip- 
ment, particularly illuminants, in high-speed photography, has had 
some interesting consequences in relation to normal cinematography. 
The possibility of using the electronic flashlight as a source for normal 
speed cinematography has been investigated, on a theoretical basis 
at least. 5 We have so far had several occasions to use it in this con- 
nection, with results that can be considered interesting and even 
promising. In general, our experience indicates that in dealing with 
live subjects intermittent flashlighting has a detrimental psychological 
effect upon organisms of a higher order. Subjects such as fish and in- 
sects seem to be affected little. People do not like it! 

The energy-storage principle used in the electronic flash unit gives 
a deceptive picture of actual power requirements in cases other than 
single, widely spaced exposures. Power-pack sizes for repetitive flash- 
ing and large light outputs rapidly reach prohibitive dimensions. 
Our experiments have therefore centered on the use of this illuminant 
for small-field work, particularly in chemistry and biology, where it's 
brilliance and low temperature permit its use close to the subject 
without discomfort or danger. We have also investigated its use as 
an illuminant for photomicrography, both by reflected and trans- 
mitted light. All these uses have been described at greater length in a 
previous paper. 3 


To Harold Edgerton and Charles Wyckoff for design and construc- 
tion of much of the equipment used in the electronic flash sequences; 
to John S. Carroll, Dorothy S. Gelatt, and Robert D. Olson for as- 
sistance in the preparation of the film "On Time and Light" and the 
gathering of material for this paper. 


(1) J. L. Boon, "The Eastman high-speed camera, Type III," /. Soc. Mot. Pict. 
Eng., vol. 43, pp. 321-327; November, 1944. 

(2) Henry M. Lester, "Continuous flash lighting An improved high-intensity 
light source for high-speed motion picture photography," J. Soc. Mot. Pict. Eng., 
vol. 45, pp. 358-370; November, 1945. 

(3) Henry M. Lester, "Electronic flash tube illumination for specialized motion 
picture photography," /. Soc. Mot. Pict. Eng., vol. 50, pp. 208-233; March, 1948. 

(4) H. E. Edgerton, "Photographic use of electronic discharge flashtubes," 
/. Opt. Soc. Amer., vol. 36, p. 390; July, 1946. 

(5) F. E. Carlson, "Flashtubes A potential illuminant for motion picture pho- 
tography," /. Soc. Mot. Pict. Eng., vol. 48, pp. 395-407; May, 1947. 

Use of High-Speed Photography 
in the Air Forces 


Summary As the conquest of flight continues, the use of high-speed 
photography by the United States Air Force is being pressed to the very 
limits of its uses and applications. By far the greatest and most varied ap- 
plication is made in the research and development center located at Wright- 
Patterson Air Force Base, Dayton, Ohio, and can be divided into two general 
categories; namely, high-speed motion pictures and high-speed still pictures. 

THE ADVANCE OF HIGH-SPEED photography has been nothing less 
than phenomenal in view of the fact that but a few short years 
ago just one camera was the basis of all this type of photography. 
This camera, a 35-mm instrument, was capable of photographing at 
the rate of 240 frames per second, was tripod-mounted and hand- 
cranked. Today this one camera has been augmented by many 
additional high-speed or ultrahigh-speed cameras. From such a 
humble beginning, the increased use of this phase of photography has 
been rapid and today is recognized as a very necessary medium in 
connection with the research and development program. Interesting 
too, has been the acceptance by engineers of this type of instrument to 
assist them in their engineering problems to give the answer where no 
other medium would suffice. 

Strange as it may seem, about ten years ago the use of high-speed 
photography in connection with research and development was a 
greater curiosity than a practical tool. However, the engineer quickly 
realized the value that this particular method of photography could 
have if properly applied. Certain changes in the procedure were 
necessary before the required answers were forthcoming. Its early 
application was mainly one of a visual study of the films as they were 
projected and the action slowed down. During the past decade high- 
speed photography has gathered components and accessories in its 
forward progress. From a very generous estimate on the part of the 
cameraman as to how fast he may have been cranking his camera or 
how accurately he may have read the camera tachometer, the methods 


82 ANDRES March 

have advanced to the point of millisecond accuracies. Today high- 
speed photography is inseparable from timing and analysis; conse- 
quently, every high-speed camera which is utilized by the Air Force is 
equipped to put a time record on the film simultaneously with photo- 
graphing or recording some specific phenomenon or action. The high- 
speed camera is no longer considered in the light of a camera; instead 
it is used as a photographic recording instrument to obtain a record 
which can later be reduced to the form of graphs or charts and more 
often than not the film record obtained by these instruments never is 
projected or sees print stage except for preservation and archival 

A brief word regarding the principal methods by which time data 
are recorded on high-speed coverage is therefore in order. 

The earliest methods of correlating time lapse with high-speed 
photography included the placing of a clock or disk driven by a 
synchronous motor in the field of the camera and simultaneously 
photographing the subject and the timing device. This method 
seriously limited the type of subject matter to that in which the clock 
and subject could be favorably disposed and illuminated. 

This led to the logical assumption that the clock should be incor- 
corporated in the camera and its image projected to the film by an 
auxiliary optical and lighting system. The Eastman-ERPI camera, 
utilizing a synchronous clock driven by a portable power supply con- 
sisting of a 200-cycle tuning fork and amplifier, exemplifies one of the 
early practical implementations of this method. The definite ad- 
vantages in facilitating analysis through direct reading of the clocks 
which are divided to YBOO of a second and readily permit interpolation 
to Yiooo of a second, still find numerous applications. 

While later developments in high-speed cameras by Eastman and 
Bell Telephone Laboratories made available increased frame speeds, 
the timing problem lagged behind these developments and these 
cameras were locally fitted with various improvised devices for placing 
a time base on the film. The earliest of these devices consisted of a 
60-cycle power supply or tuning fork of known frequency triggering, 
through a suitable electrical circuit, a spark impinging on the film. 
This method, while serving some purposes adequately, left much to be 
desired because of erratic operation. The generally poorly defined 
timing mark left on the film and the need for smaller and more ac- 
curate spacing and delineation of time increments, made necessary by 
the increasing frame speeds, proved inadequate. 




The later development employed the application of gaseous-dis- 
charge lamps such as the argon lamp. The use of small argon lamps 
as a film-marking medium was retained as they afforded a compact 
concentrated light source of high actinic value capable of responding 
with a minimum of lag to the electrical impulses and could be readily 
installed in the limited spaces available in the cameras. The major 
problem remaining in obtaining clearer delineation of timing marks 
was to modify the time- volt age curve of the pulses in order that the 

Fig. 1 Sample of synchronization mark as used in the 3-kilometer speed 
trials, showing clarity of synchronization point with respect to clock readings 
photographed on each frame. The delay or time loss from the beginning of 
signal to the beginning of recorded light on the film is negligible. 

argon lamps would achieve full intensity or extinguish as soon as 
possible after each application and removal of voltage. The charac- 
teristic sine-wave time-voltage curve of alternating-current sources 
left images on the film which varied gradually from zero density to 
full density and increased the difficulty of determining the beginning 
and end of the timing marks. This characteristic also became in- 
creasingly pronounced and objectionable with increased frame speeds. 
Taking advantage of principles developed in radar practice of 
World War II, Special Photographic Services Branch technicians de- 
veloped a pulse amplifier to convert sine-wave forms produced by 

84 ANDRES March 

various types of frequency generators into sharp square-wave pulses. 
These pulses ignite and extinguish the argon timing lights in a manner 
which produces clearly defined timing marks. A further advantage 
of such amplifier has been exploited through utilizing accurate control 
of the duration of the pulse as an additional time scale independent of 
the fundamental frequency of the timing source. 

One of the accepted methods of obtaining a time base for many 
projects, which has a wide range of application, is the 60-cycle line 
frequency where the frequency control is sufficiently accurate and 
adequate for the job. This has ground application only. Also limited 
to ground applications but highly accurate is the use of tuning forks 
of known frequency and accuracy to pulse the lamp at known inter- 
vals, thus putting a very accurate time trace on the speeding film. 
Obviously these two methods are entirely unsuitable for airborne use. 
Widely used in aircraft is the crystal-controlled chronometer of high 
accuracy which can be controlled with a minimum of effect from tem- 
perature and vibration. This method can be relied upon to give good 
performance with a high degree of accuracy. Another method fairly 
accurate and dependable, which can be airborne and meets space and 
power-supply requirements, is the electronic-oscillator circuit of the 
more stable designs. 

Born of necessity, high-speed photographic records have proved 
essential and indispensable to aeronautical research. Without this 
type of photography endless repetition of experiments and tests, with 
resultant delays, would be necessary to amass the data now disclosed 
through the medium of high-speed film analysis. 

To enumerate these applications would necessitate repeating a 
large portion of the research and development projects. Obviously, 
many of these cannot be revealed for security reasons but a few can be 


Shortly after hostilities started in World War II, it was realized that 
the flight of aircraft over wooded and isolated areas eventually would 
result in the necessity to rescue personnel who had parachuted to 
safety. The rescue problem in such cases was complicated by terrain 
unsuitable for landing a plane. The solution was to provide a means 
for human pickup by an aircraft in flight. 

Before the apparatus developed could be proved safe for general use, 


it was necessary to provide a means of measuring the instantaneous 
forces exerted due to accelerations as functions of time and horizontal 
displacement when the rescue device was picked up by the plane. 
Here, high-speed photography and subsequent analysis of the film 
provided the required data. 

Accurate marks were placed immediately behind the path of the 
rescue device and served as reference marks for horizontal displace- 
ment. Built into an Eastman high-speed camera, a timing lamp im- 
pressed known intervals of time on the edge of the film traveling at the 
rate of 1000 frames per second. This timing record, together with a 
detailed picture of the action as recorded on the film, provided a means 
of obtaining accurate time-displacement data. From these data 
accelerations and velocities could be determined. 


Since operation efficiency and thrust valves for pulse jet engines are 
functions of pulse rate and flame propagation velocity, the determina- 
tion of these quantities with precision is a vital problem of the Air 
Force. Again, high-speed photography proved equal to the task. 
A JB-2 engine, with holes 1 /4 inch in diameter and five inches apart 
drilled in the tailpipe section along its horizontal axis, was placed on a 
test stand. The flame lit up successive holes as it passed through the 
tailpipe cone and was recorded by a Fastax camera simultaneous with 
the time indications. Velocities and pulse rates could then be 

These are commonplace applications of the use of high-speed 
cameras which are daily occurrences. By far the most publicized and 
spectacular applications of high-speed photography by this group have 
been in connection with the official speed trials which have been run 
during the past two and one-half years. 

Early in 1946 the Air Force decided to make an assault on the 
world speed record which was held at that time by the British. Every- 
thing was progressing according to schedule except the timing. These 
trials must be run according to the rules and regulations set by the 
Fe*deYation Aeronautique Internationale. Investigation revealed that 
neither a method nor suitable equipment existed to time these runs 
accurately according to the high standards of accuracy officially re- 
quired. The problem was referred to technicians at Wright Field for 

86 ANDRES March 

Timing was required, but the British had discovered that verifica- 
tion of the timing equipment was also necessary. Photography was 
the answer, and high-speed photography the solution. 

In the problem of establishing a world speed record it was necessary 
to produce positive data with the required accuracy as set by the 
Fe*de>ation Aeronautique Internationale. In general these rules re- 
quired timing accuracies to at least Vwx) of a second and comparable 
accuracies in location of the airplane at points of entry and exit of an 
accurately surveyed course. 

The accurate timing was accomplished by the utilization of labora- 
tory-type tuning forks previously calibrated by the National Bureau 
of Standards at Washington, D. C. These calibrations were made 
under temperature-control conditions to enable temperature correc- 
tions in the final data. The tuning forks used were 100-cycle-per- 
second accurate to Viooo of a second. The tuning-fork oscillations 
were recorded on the film of the high-speed cameras simultaneously 
with photographed position of the contestant aircraft with respect to 
the surveyed course. A secondary timing system consisting of a 
synchronous clock was also impressed on the film photographing the 
trial plane and was used to facilitate faster analysis. These clocks 
were operated by a 200-cycle tuning fork with an appropriate elec- 
tronic system. The time indicated by the synchronous clock was 
calibrated from the 100-cycle indications on each roll of film. By this 
method the clock indications' on the film could be read in a much 
shorter time and still maintain the accuracy required by rules. The 
elapsed time for each pass could be determined by the primary timing 
or the 100-cycle-per-second indication but would require additional 
analysis time. 

The method of establishing the world's record by individually 
operating high-speed cameras was done in the following manner. As 
the plane approached the entry point of the 3-kilometer course, the 
high-speed camera at that point was started in sufficient time to 
photograph the contestant aircraft as it crossed the wire locating the 
3-kilometer entry point. Immediately following the entry of the 
course, the high-speed camera photographing the exit point was 
started so as to have both cameras in operation at the time the plane 
reached the approximate midpoint of the course. At this time a 
synchronizing point was established simultaneously on each photo- 
graphic record by applying an electrical signal to form a pulse of light 
recorded on the film of each camera. After this operation had been 




completed, the camera at the exit point would photograph the con- 
testant aircraft as it crossed the wire at the end of the 3-kilometer 
course, thus completing the operations for recording a single pass. 

In order to find the elapsed time for that particular pass the time 
indicated on the first camera as the plane entered the course was 
noted. The time of the beginning of synchronization point was then 
noted and the difference between these two readings represented 
elapsed time for approximately l /z of the course. The remaining por- 
tion of elapsed time was recorded by the second camera which was 
established by noting the time which indicated the beginning of^the 

Fig. 2 A study was made of lightweight armor-plate penetration by various 
types of 50-caliber projectiles. 

The photograph is a recording of a 50-caliber projectile trajectory just prior 
to deflection by armor plate. Two microflash units were triggered by the 
compression wave passing over two microphones and the resulting images 
were recorded on one plate. 

The projectile was of the incendiary type. Traces between the two images 
are those of the tracer compound beginning to ignite. 

same synchronization points on the second film and noting the time 
indicated upon exit of the course. The difference in these two times 
gave the remaining elapsed time of travel. The summation of the two 
elapsed times as measured by the first and second cameras was then 
added to give the total elapsed time for that particular pass. From 
this reading the speeds were calibrated to the accuracy requested. 
The identification of each pass with respect to camera location was 
made possible by slates in the field of the camera denoting pass 
number and station number. 

In accordance with the rules, four consecutive passes must be made. 
The average speed of the four passes is by rule the official speed. 

88 ANDRES March 

Keeping pace with and no less important is the application of high- 
speed still photography. Utilizing various gaseous-discharge tubes in 
combination with still cameras, remarkable results have been ob- 
tained. High-velocity projectiles have been stopped in flight with 
such clarity that the riflings are clearly discernible. (See Fig. 2.) Com- 
pression waves formed by fast-moving aircraft propellers are captured 
for analysis and study. Rupturing propellers, machine-gun malfunc- 

Fig. 3 A microflash record showing disintegration of 
frangible ammunition striking aluminum plates similar to air- 
craft covering. The ammunition was made for training of 
aerial gunners. Gunnery practice was against standard air- 
craft with suitable protection. The photographic records 
were made for the investigation of disintegrating character- 
istics of training ammunition against various types of alumi- 
num alloys. 

The microflash unit was triggered by the compression 
wave from the projectile passing over a microphone. 

tions, impact of projectiles (see Fig. 3), propeller icing, and a multitude 
of similar problems have been analyzed through this method. 

While no sure guide exists for establishing procedures or methods 
for individual problem, a careful record and file with appropriate 
descriptions are kept of each accomplishment. Regardless how 
routine or simple the process or attack may seem, many times a com- 
paratively simple approach holds the answer to a most baffling prob- 
lem. Time is too limited and qualified technicians too few to permit 
casting about for new methods when tried procedures have already 
been established and found sufficient, 


This does not mean, however, that new methods are being by- 
passed or neglected. Thinking and planning along these lines must 
continue and be encouraged. Progress in aircraft and speeds of these 
aircraft and missiles have reached fantastic proportions. Today's 
camera equipment, with modifications, is being utilized to the very 
limits of its capabilities presenting many problems. In addition, a 
definite shortage of many standard and special types of photographic 
equipment exists. 

While progress has forced the development and advancement of 
high-speed photography, high-speed photography has, in turn, im- 
plemented the rapid advancement of research and development pro- 
grams. From the simplest application, such as timing a camera 
shutter, to the study of the atom, high-speed photography has played 
a tremendous role. 

What the future holds in this respect is based to a great degree on 
supposition, but whatever the requirements, as in the past, high-speed 
photography will play an ever increasingly important function. 
Choose any field of endeavor and you have a field for the scientific 

High-speed photography has reached its place in science and en- 
gineering. While not fully exploited, its uses and applications will 
continue to increase. Quality of equipment must keep in step with 
these advancements, a proved tool it is necessary and indispensable. 

High-Speed Photography in the 
Automotive Industry 



Summary Applications of high-speed photography at the General Motors 
Proving Ground since 1938 are discussed. Equipment presently in use, in- 
cluding cameras, lighting equipment, control equipment, and accessories are 
described and illustrated. 

THE EXPERIENCE OF General Motors Proving Ground personnel 
with high-speed motion picture equipment dates back to 1938. 
Starting at that time with an Eastman Type II 16-mm camera, the 
Proving Ground has since added two more cameras and a large amount 
of associated equipment and now has the largest collection of high- 
speed photographic equipment within the General Motors Corpora- 
tion. Centrally located with respect to many of the Corporation's 
larger Divisions, the Proving Ground is within short distances of many 
plants which use our facilities for high-speed photography. 

The equipment necessarily is highly portable, since it must be used 
on the road and at many remote points on Proving Ground property, 
as well as being suitable for easy transportation by automobile to any 
Division of the Corporation. The three cameras being used are all of 
the so-called "rotating-prism" type. The original camera used by 
the Proving Ground and shown in Figs. 1 and 2 is provided with a 
motorclock for timing purposes, which is built into the camera base. 
This motorclock provides a timing image which is recorded on the 
film to the right of the picture area and within the projected frame 
area. A tuning-fork timing generator of extreme accuracy is used to 
supply the motorclock with a fixed frequency regardless of line- 
voltage and frequency fluctuations. The Western Electric 8-mm and 
16-mm cameras which we acquired during the war are similar in 
appearance (Figs. 3 and 4). An argon-lamp flasher unit is used in 
these cameras to place timing marks along the edge of the film at a 
frequency of 120 marks per second when the unit is connected to a 
60-cycle, 110-volt source. The lighting and camera-operating ac- 
ce'ssories built for this work are arranged to give portability and 
adaptability since this equipment must be used under a wide variety 


of conditions. The equipment is often operated from portable motor- 
generator supplies and, for these cases and where limited current can 
be drawn from the alternating-current circuits available at the test 
location, it is a simple ^matter to divide the electrical load among 
several sources. 

Fig. 5 shows a typical camera setup using all the camera-operating 
accessories. The Western Electric 16-mm camera is shown arranged 
to take a top view of the ejection and feeding on a 30-caliber carbine. 
Since a number of rolls of film were taken of this view, a camera volt- 
age and timing-control box was used to make camera and light opera- 
tion completely automatic upon pushing the remote-control -button 

Fig. 1 Front view of Eastman Type Fig. 2 Side view of Eastman Type II 
II high-speed camera. high-speed camera. 

shown lying below the trigger of the gun. This control ties in the 
operation of the lights to the camera and shuts the power off after the 
time required for the film to pass through the camera has elapsed. 
This control box permits the camera operator to concentrate fully 
upon the subject and makes it a simple matter to take the pictures at 
the most desirable moment. With this device the operator can con- 
trol the camera operation from any location and repeat shots for com- 
parison purposes or other reasons take only as long as is necessary to 
remove the exposed film and insert a fresh roll. Fig. 6 shows a 
schematic diagram of this control box. The Variac permits selection 
of the desired camera speed by adjusting it to the corresponding 
voltage as read on the voltmeter. The time required for the film to 





pass through the camera at that voltage is set on the timer and the 
operation is then fully controlled and initiated by the closing of the 
timing-cycle contact. 

Fig. 7 shows a camera set up to photograph the action of a milling 
machine cutting gear teeth. The box in the lower center of the 
picture supplies power to the lighting units on the stand above. These 
transformer boxes may be used in multiple to supply any number of 
lighting units. The number of lights used depends, of course, upon the 
area to be covered, the camera speed, and the type of film to be ex- 


Fig. 4 Western Electric high-speed camera with door 

posed. These units are standard General Electric Type 150 PAR/SP 
projector spotlights. Having an internally silvered reflector, these 
units concentrate all the light into a small area when used one foot 
from the subject. The power-supply box contains an autotransformer 
which allows the lights to be operated at normal line voltage for set- 
ting up and focusing, and a high-low switch on the box permits the 
application of 200 to 220 volts to the lights when the pictures are 
taken. At this higher voltage the lights operate at extreme intensi- 
ties, reaching illuminations (for three lights) of fifteen to twenty times 
that of sunlight under the most favorable conditions. A variety of 




lenses and lens extensions are used with the cameras to cover a wide 
range of field sizes and to allow some latitude as to the camera location. 
Prior to the war, the high-speed photographic equipment was only 
occasionally used since there were not many Division engineers ac- 
quainted with the time and expense savings which could be achieved 
through the use of these techniques. With the war, the work in- 
creased so that it was necessary to acquire two additional cameras and 

Fig. 5 Western Electric 16-mm camera setup used to take pictures of 
ejection on 30-caliber carbine. 

for the entire period of the war the equipment was in almost continu- 
ous use with a great increase in the application of these methods to the 
problems of the Divisions on many different types of ordnance ma- 
terial and the production machinery involved in its manufacture. 
Although the amount of work done with the cameras dropped off con- 
siderably for about two years after the end of the war, the applications 
of high-speed photography have been steadily increasing since the 
first part of 1948 as more and more engineers and production men find 
that the methods are equally helpful in the solution of problems 




involving peacetime products and production processes. A part of 
this increased use of these methods is due to a report dealing with 
high-speed photographic equipment and its applications which was 
published and distributed to the Divisions of the Corporation a little 
over a year ago. 

Unfortunately, many examples of the use of this equipment are not 
available for showing at this meeting because of the type of material 
they are concerned with and others are in the hands of the Divisions 
and cannot be located readily. Most of the work for which the 
cameras are used on the Proving Ground involves complete vehicles, 
with occasional tests on engines operating on engine dynamometers. 
Considerable work has been done on various body tests such as door- 
latch mechanisms and safety- 
glass impact tests. The action 
of springs, suspension parts and 
shock absorbers, and hydraulic 
brake lines has been studied 
through pictures taken on chassis 
dynamometers with cleats at- 
tached to the dynamometer rolls, 
and the roll speed adjusted to ex- 


Fig. 6 Schematic of camera voltage 
and timing-control box. 

cite the suspension system to its 
highest amplitude. 

The deflections of a tire striking an obstacle have been determined 
by taking 3000 pictures per second with a car speed of 40 miles per 
hour. Pictures taken of a tire striking a sharp-edged obstacle reveal 
very severe deflections. 

The action of automobile engine valves has been frequently studied. 
At higher engine speeds the surging of the valve spring becomes quite 

A study of the flow of die-cast metal from the gate of experimental 
type dies was recently made. A film of this operation shows the turbu- 
lence with which the hot metal emerges from the die gate and was 
taken to assist in the selection of a design giving the most nearly 
laminar flow from the gate on the assumption that minimum cavi- 
tation exists when the flow is the smoothest. 

The action of metal-cutting tools is of great interest to the produc- 
tion engineer. In a film showing a rather heavy cut being taken from 
cold-rolled steel on a milling machine using normal cutting speed and 
a type of tool frequently used for roughing operations, it is interesting 



to note the deflection of the tool and the apparent hesitation of the 
table as the material first strikes the tool. Higher table speeds and 
certain type tools lead to vibration or chatter of the tool which pro- 
duces a much rougher cut than that shown here. For every type of 
material there is a cutting speed and tool combination which will 
yield the most satsifactory operation and these conditions can be very 
nicely compared by the use of high-speed photography. 

Fig. 7 Arrangement of high-speed camera equipment 
for photographing a machining operation. 

High-speed photography has been found to be an extremely useful 
tool- providing great savings in time and expense in the development of 
new mechanisms and in the investigation of operating faults which 
may exist in mechanisms and processes already in use. There is every 
reason to believe that use of these methods will increase steadily as 
engineering and production personnel become better acquainted with 
the advantages of high-speed photography. 

Applications of 
High-Speed Photography 



Summary The applications of high-speed photography at the United 
States Naval Ordnance Laboratory employing oscillography, streak cameras, 
high-speed spark and discharge illumination, and rotating-prism cameras 
are described. 

THE USE OF PHOTOGRAPHY as a tool in research has increased with 
the development of new procedures for time magnification 
through high-speed photography. When these new procedures are 
introduced, they are adapted to the problem at hand, or the problem 
is so acute that a photographic method must be devised to study cer- 
tain phases of an action. 

Some of the high-speed photographic procedures for the analysis of 
the various problems encountered in this Laboratory's research and 
development activities are discussed under these general headings: 
1. oscillography; 2. streak camera; 3. intermittent illumination; and 

4. rotating-prism cameras. 


Cathode-ray-oscillograph photography is a graphic method of ob- 
taining permanent records in the analysis of amplitudes and fre- 
quencies of electrical and mechanical phenomena. High-speed 
oscillography permits the more accurate analysis of phenomena in- 
volving high frequencies. 

The Naval Ordnance Laboratory has developed and built a high- 
speed oscillograph capable of recording six traces simultaneously on 
35-mm film. It is not only capable of recording multiple channels, 
but is capable of recording high-speed transients where the function 
of time is of considerable importance. 

Basically, the instrument consists of eight RCA, Type 2BP-11, 
cathode-ray tubes mounted side by side, and six individual direct- 
current amplifiers with high gain, good high-frequency response, and 
excellent stability. Two _of the tubes provide calibrated timing 




The images of the eight tubes are transmitted to one wide-angle 
Baltar//2.3, 25-mm lens and focused onto moving 35-mm film to form 
a continuous displacement-time record of the six traces and timing 
lines. Up to 100 feet of film may be used for speeds from 4 inches per 
second to as high as 13 feet per second. When more rapid film travel 
is required and the duration of the time measurement is short, 10- 
inch strips of film may be used around the drive drum, producing up 
to 38 feet per second film speed. 

Fig. 1 Continuous print, enlarged 10 times from 35-mm film exposed in 
6-trace oscillograph. 

The Laboratory has accurately recorded signals up to 100,000 
cycles per second, which is a film displacement of O.Q05 inch from 
peak to peak at this highest film speed. Linagraph Pan, 35-mm, re- 
cording film is used when high-speed transients are involved. 

The analysis of the recording is made from a continuous photo- 
graphic print enlarged ten tunes. An extremely accurate 35-mm 
continuous printer with five or ten tunes magnification was designed 
and built for the Laboratory by Jacques Bolsey. This instrument 
prints a centimeter grid on the paper, accurate to 0.001 inch, from 
which measurements are made and which eliminates errors due to 


paper shrinkage and stretching. Paper records can be made up to 
200 feet in length on 30-centimeter stock. A typical record is shown 
in Fig. 1. 

This 6-trace oscillograph has been used in the investigation of Navy 
guns where the function of time is of greatest importance. Informa- 
tion recorded is the breaking time of the primer bridge wire, the in- 
stant of appearance of flame from the primer port holes, the begin- 
ning of recoil of the gun, and pressure-time records at three positions 
along the shell case. The time correlation of all these phenomena, 
measured from the instant the firing switch closes, is required. 

Other uses of this instrument have been made in the study of vibra- 
tions, acceleration, temperature and light measurements, and various 
types of timing measurements. 


In the study of detonation phenomena, a streak camera, involving 
the use of stationary film, rotating mirror, defining slit, accurate syn- 
chronization, and two-lens systems, was built by the Naval Ordnance 
Laboratory for the direct photography of the shock waves of various 
sizes of self-luminous charges. This is housed in a bomb shelter with 
an open tunnel-way to an outside location for the explosive charge. 
A single photograph is obtained of phenomena on a given plane as a 
function of time of each detonation from which shock- wave velocity, 
direction, shape, duration, and intensity may be calculated. The 
top writing speed of thecamera is 1.6 mm per microsecond. 

The charge is focused onto a vertical slit, variable from Viooo inch 
to */2 inch wide and 1 inch high, by interchangeable lenses of various 
focal lengths. The slit image is transmitted through a 7-inch con- 
denser lens to the rotating mirror and reflected onto a 16-inch strip 
of 35-mm film accurately placed on the interior surface of a portion of 
a cylindrical film drum. This drum is positioned to maintain a con- 
stant focal-plane distance regardless of the angle subtended by the 
rotating mirror. The image position on the film is controlled photo- 
electrically by infrared signals. 

The method of synchronization of image position, firing, safety 
factors, and mechanical shutter involved considerable development 
and improvement by the scientists of the Laboratory over other 
models of streak cameras. Because of the fact that the instrument 
would be in continuous use, every factor had to be carefully 




With the explosive charge suspended in front of the camera so that 
the plane or direction to be studied is in line with the slit, the motor 
control for the rotating mirror is started. The camera operator 
presses a firing button when the rotating mirror reaches full speed. 

By means of an electronic and relay system, the rest of the operation 
becomes automatic with the following sequence of events: 

1. The ground connection to the explosive charge is broken and 
firing connections are made. 

2. The mechanical shutter is opened, which sets up the signal in 
the photoelectric system for detonation . 





Fig. 2 Detonation camera showing optics and controls. 

3. The signal from the photoelectric cell, controlled by a light re- 
flected from the rotating mirror when in the proper photographic- 
image position, completes the firing operation. 

The schematic diagram of this unit is shown in Fig. 2 and a typical 
example in Fig. 3. 

Unusual detonation phenomena, hitherto unknown, have been 
revealed, so that the scientists working with this instrument are con- 
tinuously devising new procedures for new studies of self-luminous 

Thus far, procedures of high-speed photography as instruments for 
analysis, where the phenomena are recorded graphically, have been 
discussed. In contrast to these methods, with the advent of intense 
illumination of very short duration, actual photographs are taken 
recording a period of action that will permit the study of the behavior 


of objects traveling several times the speed of sound. Such studies 
are vital to the development of ordnance material for high velocities. 

The Naval Ordnance Laboratory now has under construction two 
aerodynamic ranges and a supersonic wind tunnel for the study of air 
flight of high-speed missiles. The ballistics ranges employ shadow- 
graph from a point source of light, and the wind tunnel a schlieren 
method of photography. 

The aerodynamic ranges will have 25 photographic stations, with 
each station equipped to photograph on 14- X 17-inch glass plates, 

Still photograph of explosive charge Streak photograph of detonation as 
with double-exposed photograph of seen through slit. Time increases to 
slit view superimposed. right. (Total time of detonation 

about 50 X 10 ~ 6 second.) 

Fig. 3 Photograph of detonation charge, using a streak detonation camera. 
Picture to the left is with mirror in stationary position. Picture to the right 
is the shock wave during detonation, showing the progress of detonation with 
time (detonation has started at bottom of view). 

the vertical and horizontal planes of the missile. At each station is 
an open-spark discharge that casts a shadow of the projectile and its 
resulting shock-wave pattern on the photographic plate. The 
diverging light traverses the range for the vertical plate and is re- 
flected to the horizontal plate by means of a mirror. When the pro- 
jectile approaches the cross-sectional area where these two diverging 
beams cross, the light is triggered photoelectrically with suitable time 
delay to center the image in this area. Open-spark illumination is 
used because of the simplicity in obtaining a fairly accurate point 
source of light and sharp cutoff in intensity. 

The value of these ballistic ranges is for the study of flight char- 
acteristics complementary to wind-tunnel models and full-scale 




missiles. The data recorded on the photographic plates will be 
analyzed for angular displacement from the bore line, and change in 
velocity as the projectile moves down range. Deceleration is deter- 
mined by an accurate timing system from station to station. This 
information is used for the study of stability, yaw, and drag over a 
portion of the flight, and, also, for correlation of differences in wind- 
tunnel models and free-flight data of full-scale projectiles. 

The study of supersonic aerodynamics by the projection of a missile 
through space is confined more to the characteristics of a projectile 
nearing completion in design and which has comparatively stable 



Fig. 4 Schematic diagram of optical arrangement for schlieren bench. 
Supersonic wind-tunnel facilities. 

flight. For velocities up to Mach number 5.18 (which is 5.18 times 
the speed of sound) the Laboratory's supersonic wind tunnel is 
used. This latter research facility is not limited by unstable flight 

This tunnel was reconstructed from the famous Kochel supersonic 
wind tunnel, used by the Germans for the development of the famous 
V-2 rockets. Its basic principle of operation is the use of a large 
sphere in which the air is reduced to a pressure of a few millimeters of 
mercury by vacuum pumps. When test operations are ready, valves 
are opened permitting the air to return through several tunnels back 
to the sphere. Each tunnel is operated separately permitting several 
tests to be performed simultaneously and independently of each other. 


The air rushes by the stationary models at a Mach number which is 
determined by the design of the control nozzles. High-speed photo- 
graphs are then taken showing the wave front of the models as if they 
were in free flight. The photographic procedure used is based on the 
principles of schlieren photography. 

The general arrangement of this schlieren method is that of a light 
from a concentrated source focused onto a slit which then becomes a 
secondary source, as shown in Fig. 4. From the slit the light goes in a 
diverging beam to one of two 500-centimeter focal-length concave 
mirrors. Parallel rays from the first mirror pass through the work- 

Fig. 5 40- X 40-centimeter tunnel No. 2 with schlieren bench in place. 

ing area of the tunnel, where the model is placed, to the second con- 
cave mirror which reflects the light in a converging beam to a focus 
where a portion of the slit image is blocked out by a knife-edge. The 
remaining part of the slit image diverges to the lens system of the 
camera and onto the ground glass, which is focused on the model in 
the air stream of the tunnel. An exceedingly clear shock-wave pat- 
tern of the model is obtained by the use of this schlieren method. 

The principle involved in the schlieren method depends upon the 
refraction of light due to the variation of optical densities in the 
vicinity of the model in the tunnel. Variations in the air density re- 
fract portions of the parallel beam farther into or away from the 

104 BEARD March 

knife-edge, resulting in a darkening or brightening of the correspond- 
ing regions of the image at the focal plane of the camera. Dark re- 
gions in the resulting photograph correspond to deviations into the 
knife-edge, while bright regions in the photograph correspond to 
deviations away from the knife-edge. Shock waves usually have 
both a dark and a bright region running parallel to each other. 

The final focusing lens may consist of single lenses or lens combina- 
tions which give the desired magnification in the final photograph. 

The cameras, of Naval Ordnance Laboratory design, are of the 
reflex principle, with 5- X 7-inch plates and commercial 35-mm 
reflex cameras. The larger camera is wired to the lighting system 
for the illumination in visual observation and when the reflex mirror 
is displaced, the light is ready for photographing. These cameras 
are interchangeable and may be replaced by high-speed motion 
picture cameras using the continuous-light source. 

The light source is a General Electric BH6 high-pressure mercury 
lamp rated at 900 watts. This lamp may be operated for continuous 
illumination or flashed intermittently. The flash duration is ap- 
proximately 4 microseconds. Photographs up to Viooo of a second are 
taken when the continuous-light source is used. The shorter exposure 
time is preferred for detailed study of small background disturbances, 
which, during longer exposures, give an integrated effect on the film. 
The schlieren setup is shown in Fig. 5. 

An example of intermittent illumination, synchronized with high- 
speed motion picture cameras, is discussed under the next heading. 


In the study of underwater ordnance, high-speed motion pictures 
with intermittent flash illumination are used in the photographing of 
exact-scale models in glass-walled tanks. These models are fired at 
various speeds up to 980 feet per second into the water and are sil- 
houetted by back lighting so that the cavitation and wave character- 
istics will form a definite line pattern outlining each phenomenon. A 
front modeling light may be used for emphasizing the shape of the 

A paper, by Anderson and Whelan, 1 pertains to the camera and 
flash units used for these studies. The paper described the sys- 
tem used at that time, but the method has now been discarded and 
completely replaced by a more flexible and efficient system. 


The light source for this system employs pulsed gaseous-discharge 
tubes and was designed to operate in synchronism with a high-speed 
motion picture camera. Repetition rates of as many as 3000 to 4000 
intermittent flashes per second have been obtained. The average 
flash duration used is about 1 microsecond. Six flashlamps are 
mounted in one single-parabolic reflector, six feet in length and run- 
ning the length of the back of the model tank. The control equip- 
ment, designed by the Naval Ordnance Laboratory, permits the inter- 
mittent flashing of each lamp independently or in any sequence or 
combination. This flexibility is highly desired where progressive 
sequence lighting is required to "follow" the action across the tank or 
where special lighting is needed for the elimination of unwanted 
shadows. A diffusion screen is used over the face of the reflector. 


High-speed 16-mm motion picture cameras of the revolving-prism 
type are used and are equipped with a new electromagnetic pickup, 
without any mechanical contact with moving parts of the camera, for 
absolute synchronization of the camera with the flashlamps. The 
complex synchronization and lighting system is so designed as to be 
used with other types of high-speed cameras and varieties of flash- 
lamps or open-spark circuits. 

The revolving-prism cameras were selected for use after compari- 
sons were made with other types of high-speed motion picture 
cameras. The rotating prism satisfactorily eliminated blurring of 
the image which was noticeable when cameras not employing this 
system were tested. Image displacement for still objects, when there 
is no compensation for film motion, is 0.028 mm at 1 microsecond 
flash duration at a camera speed of 4000 frames per second on 16-mm 

Other uses of cameras with rotating prisms are as all-purpose 
cameras at the Naval Ordnance Laboratory, when the magnification 
of time is essential. High-speed motion picture photography with 
these cameras has become so generally required that they are assigned 
to staff photographers and engineers as regular equipment. Some 
field tests require the use of several cameras on one assignment. 

The general use of these cameras is quite similar to those applica- 
tions in industry where impact, timing, or erratic operations, in- 
volving moving mechanical parts, are studied, and daylight illumina- 
tion or portable floodlights are used. This type of usage, therefore, 

106 BEARD 

would be similar to those described by others in this symposium. 
This paper has been presented to acquaint the members of the 
Society with a general idea of the various types of high-speed pho- 
tography undertaken at the Naval Ordnance Laboratory and, there- 
fore, the paper could not be confined to a detailed study of any one 
operation. The photographic methods included represent the efforts 
of numerous engineers and scientists from various divisions of this 
Laboratory, and the author is indebted to those people for the above 
information. Particular indebtedness is to G. E. Beyer and J. L. 
Jones, Mechanical Evaluation Division, Technical Evaluation De- 
partment on oscillography; S. J. Jacobs, Explosives Division, Re- 
search Department on the streak camera ; members of the Mechanics 
Division, Research Department on aerodynamics; and members of 
the Hydrodynamics Subdivision, Research Department on under- 
water ordnance. 


(1) R. A. Anderson and W. T. Whelan, "High-speed motion pictures with 
synchronized multiflash lighting," J. Soc. Mot. Pict. Eng., vol. 50, pp. 199-208; 
March, 1948. 

Control Unit for Operation of 
High-Speed Cameras 



Summary An automatic time control for the Fastax camera doubles the 
picture-taking speed of this camera, controls the film speed from low to 
high, automatically controls the camera in' synchronism with an event to be 
photographed, allows for remote-control operation of the event and the 
camera, and prevents tearing of film as the camera is started. It is simple 
in operation, extremely accurate, and completely portable. 

THE 16-MM Western Electric Fastax camera as originally developed 
by the Bell Telephone Laboratories could attain a maximum 
speed of 4000 frames per second. Later it became necessary to speed 
up the action of the camera so that motion picture studies could be 
made of detonations, vibrations of propellers, supersonic studies of 
airflow over wing tips, and many other ultrafast actions too numerous 
to mention. 

Bell Laboratories determined that it was possible to double the 
speed of the camera by applying 280 volts directly to the camera 
motors. However, when starting the camera at this speed, the film 
would strip at the sprocket holes. Therefore, a 70-millisecond delay 
circuit was interposed so that the camera could start at 130 volts, 
run at a speed of 4000 frames per second for 70 milliseconds and, by 
jumping the voltage up to 280 volts, bring the camera to the full 
speed of 7500 frames per second. This peculiar jumping action led 
to the name of the "goose" for the Model J-410 control unit. 

With the advent of these high speeds it became increasingly more 
difficult for the operator to synchronize the camera to the event which 
was being photographed because 100 feet of film travels through the 
camera in 8 /i of a second. 

The Industrial Timer Corporation was approached as to the pos- 
sibility of designing a circuit which would enable the operator to syn- 
chronize, by means of time, any action-to-camera requirement which 
could be encountered in the laboratory or in the field. It developed 
the Model J-410 control unit (the "goose") which not only took care 





of the synchronization of the camera to the event, but also included 
the 70-millisecond time-delay circuit previously described. 

Fig. 1 is a photograph of the front panel unit showing timer 1 
(camera timer), timer 2 (event timer), voltmeter (for voltage-to- 
speed readings), variable-transformer knob (for camera- voltage ad- 
justment), and all the switches and receptacles necessary for the ef- 
ficient operation of the high-speed camera. 

The camera timer controls the period that the camera is operated 
and cuts off the camera voltage at the end of the preset time cycle. 

Fig. 1 

This timing cutoff is important to prevent damage to the camera, 
which must not be run at high speeds after the film has passed through 
the aperture. The camera-time period is preset according to a speed- 
to-voltage-to-time chart based on the operating characteristics of the 
Fastax camera (Fig. 2). 

The event timer sets up the time necessary for the event that is 
being studied to prestart or poststart the camera. If the event can 
be started or stopped by an electrical switch, the contacts on the 
event timer can be used efficiently for this purpose. These contacts 
come out to a receptacle on the front panel labeled "event." Timer 
3, or the 70-millisecond delay timer is installed inside the housing. 

A remote-control receptacle is provided to enable the operator to 




set up his equipment and operate the camera from any distant point 
when it is necessary to do so for safety reasons. 

Because the Model J-410 is u^ed in both tropic and arctic areas, a 
100-watt heat lamp is installed within the case which acts to keep 
the unit dry in the tropics and warm in the arctic regions. 


6 MM 16 MM. 35 MM 



1.0 .8 

14000 7000 

12000 6000 

10000 5000 4000- 
8000 4000 


eooo 3000 

4000 2000 

160 200 24O 2*0 


Fig. 2 Settings based on use of 100- 
foot film rolls only. 

The "goose" can also be used efficiently with the Eastman high- 
speed camera by setting the voltage for the camera at 120 volts and 
adjusting the camera as normally used. 

The housing is finished in baked black crystal enamel and measures 
21 X 10V2 X 15 inches. 

Lenses for High-Speed 
Motion Picture Cameras 



Summary A description is given of the objective lenses which are avail- 
able for use on the Western Electric Fastax cameras. The range of focal 
lengths is from 35 mm at an aperture off/2.0 to 380 mm at//5.6. A mount- 
ing of the bayonet type is provided for all lenses to make them readily inter- 
changeable on the camera. A reflex viewfinder for the 35-mm camera is 

THE GENERAL CHARACTERISTICS of the cameras designed for high- 
speed photography at the Bell Telephone Laboratories have been 
described in a series of papers published in this JOURNAL. J ~ 3 

There are three models of these cameras, one for each of the three 
standard size films: 8 mm, 16 mm, and 35 mm. The cameras have 
one important feature in common, that of a continuous film motion 
which in turn is compensated by a glass block or prism located between 
the lens and the film. This compensation is effected by rotating the 
prism at the exact speed required to move the image down across the 
aperture in synchronism with the motion of the film. 

The optical effect of the prism in these cameras is an important 
factor in the selection of the lenses that are to be used. In the matter 
of shape, all the prisms are optically similar; all are glass plates with 
plane-parallel surfaces. The result of placing such a plate behind the 
lens in a camera is a familiar one in optical design. 4 In general, the 
effect is that of a negative lens. The back focus of the camera lens is 
increased. The plane-parallel plate introduces aberrations which 
change the pattern of image formation at the focal plane, and this 
change usually means a loss in image quality. The amount of the 
aberrations thus introduced is directly proportional to the thickness 
of the glass plate. It varies also with the aperture of the camera lens 
and the angular field covered by the lens on the film. , 

Therefore each type of lens used on a prism camera presents an 
individual problem. The camera lens is a complex system of refracting 
elements designed to give a well-defined image at the film plane. It is 


possible to modify one or more of these elements to correct for the 
aberrations of the plane-parallel plate and thus restore the image to its 
original quality. The correction is usually computed for the normal 
position of the prism, when the parallel faces are at right angles to the 
lens axis. This method of correction provides a practical solution of 
the rotating-prism problem. However, it is not an ideal solution ; it 
does not take into account the complex changes in prism aberrations 
which are produced by rotation of the prism surfaces during the ex- 
posure cycle. 5 The results must be checked by test films exposed 
under actual operating conditions. This has proved to be a satis- 
factory and very useful method in the development of a series of 
lenses for the Fastax cameras. 

Like all new projects involving complicated machines, the high- 
speed camera development proceeded gradually to its present state. 
In the early days the three cameras could not be considered together, 
with a set of drawings and a list of lens requirements attached. At 
that time there were only two cameras, the 8-mm and the 16-mm 
sizes. These two cameras have prisms of identical thickness, although 
the prism for the 8-mm camera is an octagon and that for the 16-mm 
camera is a square in cross section. A lens of 51-mm focus with an 
aperture ratio of f/2 was furnished for both these outfits and later 
adopted as standard equipment. The angular field for these film 
dimensions is small, 5.5 degrees on the 8-mm camera and 11.75 de- 
grees on the 16-mm camera. Resolving-power tests show 28 lines per 
millimeter on the 16-mm camera at //2, with an increase to 40 lines 
per millimeter at the center of the picture. At //3.5, 40 lines per 
millimeter are resolved throughout the film area. These results 
apply to Super XX film and Kodachrome. 

With the introduction of the 35-mm camera a few years later the 
lens problem became more complicated. This camera was designed 
to photograph an area 40 degrees in width and 12 degrees in height 
on a half frame of 35-mm film. The focal length required is 35 mm. 
An aperture of f/2 was specified, new prism dimensions were in- 
volved, and at the same time it was decided that lenses should be 
made interchangeable in all three cameras. Since the difference in 
prism thickness involved was only 0.008 inch, it seemed best to make 
a new 35-mm f/2 objective corrected for the larger prism and the 
widest angle required, assuming that any lens with definition and 
covering power sufficient for the wide-angle camera would be satis- 
factory on the smaller sizes. 




The first results of this new 
design were not entirely satisfac- 
tory. Lenses from the first pro- 
duction lots were used on wide- 
angle cameras and although many 
useful films were produced with 
them, the resolution at the mar- 
gin of the film was less than ex- 
pected. It took more time to 
find the cause of the trouble 
overcorrection of field curvature 
and this has been eliminated 
Fig. 1 35-mm lens assembly. within the past year. The 

wide-angle camera will now resolve 28 lines per millimeter, and 
reaches 40 lines per millimeter in the central part of the film. This 
35-mm, f/2 camera lens is a 6-element design of the Biotar type of 
construction. It has been corrected for prism aberrations by the 
method just outlined. In respect to distortion there is still room for 
improvement; 1 per cent is the amount remaining at the margin of 
the picture area. All air-glass surfaces of the lens are fluoride- 
coated to minimize reflection effects, and this coating is applied on all 
lenses for these high-speed cameras. 

The mounting and mechanical design of the 35-mm lens is shown 
in Figs. 1 and 2. The slotted cylindrical bearing surface of the 

Fig. 2 Reflex finder and lens for 35-mm camera. 




mount is 1.775 inches in diameter and 0.320 inch long. This arrange- 
ment is designed to fit the bayonet socket of the cameras, and the 
same lock ring is used for all the lenses of the series. The inner de- 
tails of the lens mount and focusing sleeve permit a par-focal adjust- 
ment for any one of the three cameras. 

Fig 3 Complete series of camera lenses. 

Fig. 2'also shows a reflex finder made for the 35-mm camera. With 
this device the photographer can observe and check all details of his 
subject in setting up the camera, lights, background, and so forth. 
The camera lens fits on the finder for this purpose, and therefore it has 
two scales on the focusing ring. The red scale applies to its use as a 

114 COOK March 

finder lens. When the setup is complete, the finder is removed and 
the lens replaced on the camera. The white scale is then used, its 
figures giving the correct distances from the camera position. 

As the high-speed cameras were developed and applied in many 
fields of research, the need for other types of lenses became apparent. 
Long-focal lengths are required to get details of distant objects, and 
these longer lenses were designed one by one as they were needed when 
the cameras were applied to various new problems in motion analysis. 
In considering these long-focus lenses, it is important to note that 
their aperature ratio is less than//2. The field of view requirement, 
as shown in Table I, is also small. It follows that the effects of the 
prism surfaces on image quality are less disturbing as the focal length 
increases, because the angles of incidence of the separate light rays 
are smaller. The design of these long-focus lenses is simplified 


















































Table I gives a list of the principal dimensions of the whole series of 
lenses produced for high-speed cameras. For the 75-mm and 101-mm 
focal lengths a construction of the Tessar type has been used. The 
last three members of the series 150-mm, 254-mm, and 380-mm 
focus are of telephoto design, with the characteristic short back 
focus to reduce the over-all length to reasonable limits. Fig. 3 
shows some details of the mountings and the relative sizes of all 
lenses of the series. 

The final test of any photographic lens is the result obtained on the 
film in the camera, under operating conditions. In general the resolu- 
tion of these high-speed cameras is 28 lines per millimeter; 40 lines 


per millimeter are resolved in some cases. This is not a high figure 
for ordinary motion picture standards and it certainly will be im- 
proved in the future. 

In a rotating-prism camera the requirements of synchronization at 
high speed are opposed to those of resolution in some respects. 
Synchronization demands one set of conditions, high resolving power 
a different one. Both optical and mechanical problems are involved 
here, and they must be considered together in their effect on the per- 
formance of the high-speed cameras. 


(1) W. Herriott, "High-speed motion picture photography applied to design 
of telephone apparatus," /. Soc.. Mot. Pict. Eng., vol. 30, pp. 30-38; January, 

(2) Howard J. Smith, "8000 pictures per second," /. Soc. Mot. Pict. Eng., vol. 
45, pp. 171-184; September, 1945. 

(3) John H. Waddell, "A wide angle 35-mm high-speed motion picture cam- 
era," J. Soc. Mot. Pict. Eng., vol. 46, pp. 87-103; February, 1946. 

(4) I. C. Gardner, Scientific Papers of the Bureau of Standards, No. 550, p. 
160; May, 1927. 

(5) John Kudar, "Optical problems of the image formation in high-speed motion 
picture cameras," /. Soc. Mot. Pict. Eng., vol. 47, pp. 400-403; November, 1946. 

High-Speed Photographic System 
Using Electronic Flash Lighting* 



Summary By way of summary it may be said that a system of high-speed 
photography has been developed which combines the desirable features of 
rotating prism-type motion picture cameras with those of electronic flash 
lighting. This has resulted in a system which delivers extremely high over-all 
definition and incorporates operating flexibility. Special lighting effects are 
made available by means of all-electronic interlacing equipment. The system 
can be operated in its present form at frame rates as high as 8000 frames per 
second without reduction of the illumination available per flash. An effective 
exposure time which never exceeds several microseconds is realized, and 
interchangeability of various commercial flashlamps is possible without 
modification of the electronic circuits. 


AN ELECTRONIC flash-illuminated system of high-speed photog- 
raphy has been developed at the Naval Ordnance Laboratory 
for investigation of short-duration, nonrepetitive events. The 
Laboratory setup has a number of features in common with other 
systems which have been reported previously in the literature. lf 2 
Certain specific differences exist, however, which justify this 

The system consists of a rotating prism-type high-speed camera 
with which one or more electronic flashlamp units are synchronized 
so that the individual frame exposures are made at the instant at 
which the prism shutter is "wide open." Additional gating, sequenc- 
ing, and interlacing equipment are included which provide picture- 
taking flexibility not reported heretofore. 

Before entering upon a detailed description of the system, it is 
necessary to point out that it is not presented as an end product, but 
merely represents a pilot model of a larger, more complete unit which 
is being developed at present. However, it has been in service long 
enough, and a sufficient number of test scenes have been taken with 
the apparatus to prove that it can be used readily as a working unit 
in its present form. 

* This work was supported by the Office of Naval Research. 



A discussion of the system should be prefaced by a consideration of 
what features it makes available to the user. The most important 
characteristic of electronic flash illumination, that of an extremely 
short exposure time with its corresponding motion-arresting capabili- 
ties, is featured in the system. The light sources to be described per- 
mit an effective exposure time which varies from 1 to 3 microseconds. 
The exact duration of the flash is dependent upon two factors : (a) 
the particular type of flashlamp being used and (b) the electrical input 
in a specific setup. As implied by this remark, a variety of com- 
mercial flashlamps can be used interchangeably in this system, and the 
electrical power input can be varied to suit the needs of the operator. 
Any camera speed up to 8000 frames per second can be accommodated 
readily. Adaptation to any of the various commercial high-speed 
cameras, whether they are constructed with or without compensating 
prisms, seems readily achievable although only the Eastman Type 
III is in use at present. Sufficient light output is available for most 
scenes so that adequate exposure is not a serious problem until frame 
rates in excess of 4000 per second are reached. Above this speed it is 
necessary to increase the number of lamp units per scene over that 
which slower frame rates might demand in order to distribute the 
burden and prevent lamp overload and its resulting lamp destruction. 
Provision is made for the simultaneous, synchronous operation of as 
many as six lamps from a single camera, and the entire system is 
operated from a single control panel. The duration of any specific 
lamp's working interval, as well as the sequence of operation of the 
individual lamp, is adjustable and self-indicating at this panel. 

An important feature is the availability of half-frame-rate operation 
of any specific lamp or group of lamps. If the instantaneous camera 
speed is, say, 3000 frames per second, then this convenience makes it 
possible to operate any desired lamp at a corresponding instantaneous 
rate of 1500 flashes per second. This method is referred to in this 
article as "half -frame-rate" operation and is meant to refer to the 
flash rate of the lamp and not the picture size. For normal projec- 
tion, this half-frame-rate operation is not desirable, since, if part of 
the lamps are operated at half -frame rate of 1500 frames per second, 
and the rest are operated at the full camera speed of 3000 frames per 
second, the only effect one would notice on the projected motion 
picture would be an annoying flicker. However, for frame-by-frame 
analysis purposes, alternate frames which can be provided with two 

118 WHELAN March 

different types of lighting represent potentially twice as much experi- 
mental data per reel of film since such a method of photography might 
permit the simultaneous recording of two entirely different aspects of 
the action. For example, it is often necessary in the photography of 
objects entering water to study the event not only by means of re- 
flected or " conventional" lighting, but also by means of transmitted or 
"silhouette" lighting. With two banks of lamps properly disposed, 
both lighting effects could be interlaced on a single reel of film. 

In addition to differential lighting effects, this film-sharing prin- 
ciple can be used in such a way that higher frame rates can be accom- 
modated by a bank of flashlamps. As mentioned previously, the 
tolerable number of flashes per second at a given input per flash is 
essentially limited by the flashlamp at the present time, since under 
conditions of constant energy input per flash the total average power 
input to a lamp is directly proportional to the frequency of flashes, 
and a maximum average power input exists for any type of flashlamp. 
Because of this power-handling limitation of the flashlamps, normally 
one is forced either to reduce the "per flash" energy applied to the 
lamp or to restrict the permissible total number of flashes; i.e., the 
duration of the operating cycle must be reduced. Neither one of 
these possibilities is desirable from the photographic standpoint. This 
difficulty has been overcome in the present system by use of the half- 
frame-rate provision. Two such half-frame-rate channels operate on 
alternate frames and permit individual light sources to operate at 
one half of the instantaneous camera-frame rate. In this manner, 
even for camera speeds as high as 8000 frames per second, the flash- 
lamps can be operated on a 4000-frame basis. Naturally, the two 
groups of lamps which are thus duplexed must be so placed that they 
effect the same subject illumination. The principle of film-sharing 
can be extended to provide for even higher camera-frame rates, but 
this has not been done in our system. However, a modification which 
permits extension of a system to 16,000 frames per second will be re- 
ported in the near future. 


A block diagram of the electrical system is shown in Fig. 1. Since 
the individual items will be described in some detail in the following 
pages, only a functional description of each is given here. 

The entire system is designed around the camera itself, and as it 
has been remarked, the system described here is adapted to the 




Eastman Type III camera. Fortunately, when camera types are 
changed, the necessary modifications are relatively minor. The 
Eastman camera was chosen since it admitted of ready adaptation to 
a multiple-flash system, and although its maximum speed was in the 
neighborhood of 3000 frames per second, this speed was found en- 
tirely adequate for the specific problems under study. Referring to 
Fig. 1, the unit described as "master control" is a panel which con- 
tains two main switches and a group of signal lights. The first switch, 













Fig. 1 Block diagram of the system. 

when closed, merely permits a controlled five-minute preheat of all 
the electron-tube cathodes in the system. At the end of this interval, 
it is possible to throw the second "event" switch and thus initiate the 
picture-taking cycle. Under normal conditions of operation, these 
are the only two actions which are required to take a picture other 
than the usual photographic duties. These include the camera-load- 
ing, setting of the "prerun" footage to be passed through the camera 
while it comes up to operating speed, setting the maximum camera 
speed desired, placing the lights in appropriate positions, focusing, 

120 WHELAN March 

making aperture adjustments, and setting of the sequence and inter- 
lacing equipment. Normally, if a series of shots are scheduled which 
do not require appreciable scene changes, a run can be set up and taken 
in a matter of minutes. 

When the event switch is thrown, the direct-current power supplies 
which feed the system and the camera motor are energized. The 
camera immediately accelerates, and a small alternating voltage, 
whose frequency is proportional to the optical shutter speed, begins 
to be fed to the "synchronizing shaper and divider" unit at the control 
position. In this unit, this small voltage, called the synchronizing 
signal, is transformed into an abrupt trigger pulse whose leading edge 
coincides with the instant of the wide-open shutter. A divider circuit 
supplies two other outputs of synchronizing pulses which correspond 
in time to the wide-open shutter positions of alternate frames; i.e., if 
the instantaneous camera speed is N frames per second, synchro- 
nizing pulses appear at one of the outputs and occur at a rate of N/2 
frames per second. These pulses coincide with the wide-open shutter 
positions on, say, the even-numbered frames, while at the other out- 
put pulses also occur at N/2 frames per second, but these coincide in 
time to the open shutter on the odd-numbered frames. 

The synchronizing pulses are then fed to the ' 'line-amplifier and 
gate" unit where they are subsequently amplified without phase 
change and passed on to six channels that control specific flashlamps. 
No progress of the pulse through these channels would occur unless 
an outside influence acted at this point, since an electrical gate oper- 
ates in each of the six channels and normally blocks the passage of any 
signals. The camera, itself, provides the outside influence to unblock 
the gate circuits at the proper instant and permits the synchronizing 
pulse to advance. This unblocking signal is^enerated in the Eastman 
camera by the closing of a microswitch at a predetermined instant. 
The unblocking information is relayed over the "start-picture" signal 
line. Normally, the unblocking "start-picture" signal is delayed 
until the camera approaches operating speed. When this unblocking 
information arrives at the "sequence and interlacer" unit, an elec- 
trical signal is developed for triggering the event to be photographed. 

In the sequence and interlacer unit, in addition to the event trigger 
signal, the sequence and duration of the unblocking signals which 
close the individual electronic gates in the 6-channel line amplifier are 
determined. Six cascade unblocking periods whose durations are 
variable from zero to 3 seconds in length are available; i.e., the second 




timed interval commences immediately upon the cessation of the 
first, the third commences immediately upon the cessation of the 
second, and so forth, until the six timed intervals have occurred. This 
arrangement permits an operator to set up a sequence of lighting in 
which the area illuminated will change with time. In this way, pro- 
gressive lighting can be preset to "follow" a relatively slow-moving 
object across a large field of view and so the operating time of the 

Fig. 2 A portion of the equipment showing power supply 
(right), control panel (left), Eastman camera, and two 
pulse modulators with their associated flashlamps. 

flashlamps is restricted to the barest minimum. Thus by proper 
choice of the interconnection of the three available synchronous 
signals, and the six variable unblocking gate signals, the operator can 
choose from an extremely large number of possible lighting effects. 

Once the synchronous signal has passed through an appropriate 
"closed" line-amplifier channel, it is fed over cables to one of the six 
remote units referred to as the "high-level pulse modulators." These 

122 WHELAN March 

units are pulse generators that transform the direct-current energy 
from the main power supply into abrupt high-voltage surges of less 
than a microsecond duration which pass down cables to the individual 
gaseous flash units where they supply the electrical input energy of the 
light sources. The instant at which the discharge pulse, and hence the 
light output, takes place is precisely controlled by the arrival of the 
synchronizing pulse at the modulator unit. As mentioned previously, 
this instant is made to coincide with the wide-open position of the 
optical shutter in the camera. 

In Fig. 2 is shown a partial setup of this multiple-flash high-speed 
photographic system. The rack on the left is the control panel which 
contains the units just described, and the rack on the right contains 
the direct-current power supply. At the extreme right is a pair of 
remote pulse modulators, and in the foreground is a pair of mounted 
commercial flashlamps. The Eastman camera stands behind the 
modulators on its own rigid tripod. 


Certain basic limitations exist in the system outlined above and 
should be noted here, (a) Only one camera can be used with it, unless 
one resorts to expensive and seemingly impractical "ganging" of the 
rotating systems of multiple cameras ; (b) a single ' high- voltage 
direct-current power supply feeds, six lamps, and thus any failure in 
the power supply removes all six lamps from service; (c) overload 
protection is provided in the low-voltage primary circuit of the power 
supply which means that a high-voltage fault at any point in the 
high-voltage system results in the complete failure of all lamps; 
(d) the high-voltage power supply requires a 3-phase primary power 
line of considerable capacity which restricts the number of possible 
locations where the equipment may be operated ; and (e) the weight of 
the system is of the order of 2000 pounds which further restricts its 


The pilot system described was built of conventional electronic 
components and hardware wherever possible. The high-voltage 
power supply and the control units are housed in a pair of enclosed 
transmitter cabinets, each of which provides seven feet of usable panel 
height. A third cabinet is required if operation below frame speeds of 
1000 pictures per second is desired. The individual pulse-modulator 


units are built into relatively small steel boxes and are normally 
located remote from the control panel and power supplies in order to 
keep the line between the actual flashlamp unit and the pulse modu- 
lator short. All cables are made up in standard interconnecting 
lengths, and are of seven general types; (a) low-level pulse cables 
which connect the interlacer panel to the remote pulse-modulator units, 
(b) high-level pulse cables which connect modulators to lamps, (c) high- 
voltage direct-current power lines which carry direct-current power 
into the pulse-modulator units, (d) camera motor power and control 
lines, (e) camera-synchronizing lines, (f) conventional single-phase, 
low-voltage power lines, (g) heavy-duty, 3-phase power lines. All 
lines which carry synchronizing or high-power pulse signals are of the 
shielded coaxial type which prevents unwanted interaction of units. 
Successful photography has been accomplished with modulators, 
lamps, and cameras placed within an explosion chamber which was 
located over 150 feet away from the control panel. Normally, it is 
desirable to keep the cable length between the modulator and flash- 
lamp less than 50 feet. For a problem where the modulator-to-lamp 
distance could be kept constant, it would be more desirable to house 
the pulse modulators in a single transmitter-style cabinet. 

Since a variety of flashlamp light sources have been used in the past, 
various types of lamp housings and reflectors have been used. Perhaps 
the most convenient lamp units were realized when the General Elec- 
tric Company's Type FT-125 lamp was used. This lamp is built into a 
"sealed-beam" headlight unit and thus provides its own reflector. A 
very simple mounting was provided by fixing the lamp unit inside 
a small, commercially available steel cabinet. Lamps of the long, 
slender geometry were usually mounted in hastily contrived cylin- 
drical reflectors which are parabolic in cross section. For silhouette 
lighting effects, all of the light sources were provided with diffusers in 
order to provide a background of uniform illumination. 

A serious problem of the electromechanical design was that of pro- 
viding safe high-voltage connectors for the direct-current power 
cables, as well as the high-voltage, heavy-current cables which con- 
nect modulators and lamps. At the time of the development, these 
difficulties were eliminated by manufacturing special connectors. 
However, a more practical solution would be to design the cables for 
use with standard high- voltage X-ray fittings or their equivalent. 
The usual door interlocks and "dead" front panels were provided for 
personnel safety. 

124 WHELAN March 


A thorough discussion of all of the electrical and electronic details 
would require an undue amount of space and would be inappropriate 
at this time. For this reason, the present discussion may seem 
sketchy, but the References should provide the interested reader with 
whatever information he may desire. 

A common element in any electronic flash-lighting system is the one 
which forms and controls the short burst of electrical energy which is 
transferred to the gaseous-discharge lamp at the instant light output 
is desired. The simplest method of accomplishing this task is to 
charge an electrostatic capacitor to a voltage somewhat below that 
required to cause electrical breakdown of the gas column. Then at the 
appropriate instant, an external agent is brought into play which 
momentarily destroys the insulating property of the gas, and thus 
causes electrical breakdown within the gas column. In a matter of a 
few millionths of a second, this gas column becomes a highly con- 
ducting load on the capacitor and quickly discharges the stored charge 
from its plates. Normal electrodynamic-gas processes occur which 
result in the radiation of a large amount of energy, much of which is 
photographically active. 

The foregoing method of triggering is usually used in small, single- 
flash systems. The multiple-flash system described here differs some- 
what in its mode of operation from this method in that the capacitor 
is normally charged to a voltage considerably higher than that re- 
quired to cause spontaneous breakdown of the gas column. As a con- 
sequence, the capacitor must not be connected across the lamp load 
until the very instant at which the light output is desired. Such a 
requirement demands a switch which is capable of holding off ex- 
tremely high voltages until the proper instant, and which can be 
closed in a matter of microseconds by simple electrical means. In 
addition, the switch must be capable of carrying the capacitor-dis- 
charge current repeatedly without incident or excessive loss of elec- 
trical energy. A switch which meets these requirements exists in the 
hydrogen thyratron, a tube which was developed during the last war 
for certain radar problems. 

These thyratrons are made in a variety of sizes, and the particular 
one used in the pulse modulator described here is designated as the 
Type 5C22. It is rated at a peak current of 325 amperes, while the 
normal discharge current through the flashlamp is of the order of 1500 
amperes. Even though the ratings are thus exceeded each time a 


flash occurs, the short duration of the discharge-current pulse (about 
0.75 microsecond for the usual circuit constants) and the fact that the 
longest picture-taking runs of less than 3 seconds in length doubt- 
less account for the satisfactory thyratron life experienced. Ulti- 
mate failure is caused by hydrogen "cleanup" in the internal tube 
structure. The circuit diagram of the modulator is shown in Fig. 3. 
Vi is the 5C22 hydrogen thyratron, V 2 is the flashlamp, and V 3 is a 
clamping diode. Ci is the impulse energy-storage network, and LI is 
the iron-cored recharge-control inductor. The purpose of L 2 is to 
permit the recharge current to by-pass the flashlamp. G is a protec- 
tive gap which prevents overvoltage operation in case of circuit 
maladjustment. L 3 , L 4 , and Cz constitute a pulse-coupling network in 
the grid of the switching thyratron. This circuit is a conventional 
inductive-charging modulator, and its operation is well covered else- 
where. 3 - 4 The singular features of it worthy of note are: (a) it is a 

Fig. 3 Simplified schematic diagram of the pulse- 
modulator unit. 

voltage-doubling circuit; i.e., under conditions of proper adjustment, 
the energy-storage capacitor will be charged every cycle to a peak 
voltage approximately twice that of the direct-current source, (b) the 
over-all charging efficiency (defined as the ratio of the capacitor 
energy made available for flash to the energy drawn from the direct- 
current source) is high, (c) it readily adjusts itself to various flash 
rates and hence various camera speeds. 

If one has been accustomed to thinking in terms of single-flash 
electronic illumination, the matter of charging efficiency might appear 
at first to be of secondary importance, but for multiple-flash operation 
during which the recharge cycle may occur thousands of times per 
second, charging efficiency must be high in order to prevent undue 
loss of electrical energy. For this reason, simple resistance-controlled 
charging is highly undesirable, since under optimum conditions of 
adjustment this method results in an efficiency ratio of only 0.5. On 
the other hand, inductive charging approaches an efficiency ratio of 

126 WHELAJST March 

1.0 as the losses in the charge circuit are made smaller. These losses 
include the iron and copper losses of the choke LI, the dielectric losses 
of the storage capacitor Ci, the plate-circuit losses of the clamping 
diode F 3 , and the internal power-supply losses. This clamping 
diode F 3 , which is indicated in Fig. 3, is a necessary evil which cannot 
be avoided if the recurrent flash rate of the system falls below one 
half the natural resonant frequency of the charging choke LI, and the 
storage capacitor C\. Its purpose is to prevent the overvoltaged 
storage capacitor from discharging back into the power supply. In 
the present system, the size of the parameters is such that this diode 
is not required unless the flash rate falls below 1000 per second. 
Housing for the six diode-clamping units requires a third transmitter- 
type cabinet if such low-frequency operation is desired. 

It is of interest to consider the power requirements of the system 
under conditions when the frame speed is considerable and when the 
high-efficiency charging circuit is used. The average direct-current 
power required by a single-pulse-modulator unit is computed readily 
if it is assumed that the entire block of electrostatically stored energy 
in the capacitor is completely transferred from the capacitor to the 
discharge circuit during each discharge cycle, and must be re-estab- 
lished at the expense of the direct-current power supply. This reason- 
ing leads to the expression for the average power given below. 


where P = the average direct-current load demand in watts if the 
other terms are defined as follows : 

n = number of discharge cycles per second (flash rate) 

C = effective capacitance of the storage capacitor in farads 

V = maximum voltage (in volts) to which capacitor is charged each cycle 

ij = over-all efficiency of the charging circuit expressed as a decimal. 

In the system described, the nominal storage capacitor used is a 
0.05-microfarad capacitor, the crest capacitor voltage is of the order of 
10 kilovolts, and the charging-circuit efficiency can be expected to 
run about 0.9. Under these conditions, a single lamp flashing in syn- 
chronism with a high-speed camera, running 2500 frames per second, 
would require 3.14 kilowatts from a direct-current source of slightly 
more than 5000 volts. Six such lamps would require a total average 
power of over 18 kilowatts. Assuming a direct-current power-supply 
efficiency of 0.9, it is seen that the average alternating-current power- 
line load is more than 20 kilowatts. Such large power demands, even 


though they may exist for but a few seconds, require rather careful 
planning of the power supply itself. It would be extremely poor en- 
gineering to attempt to build a 5-kilovolt, 20-kilowatt, single-phase 
rectifier set which would be required to supply smooth, ripple-free 
direct current to the load. For this reason, a full-wave, 3-phase bridge 
rectifier set was used. It operates from either 220- volt or 440- volt, 
3-phase circuits, and is protected against either alternating- or direct- 
current faults by fast-acting air-circuit breakers. 

The requirement that the camera must develop an electrical syn- 
chronizing signal, which was proportional to the rotating-prism 
shutter speed, caused some difficulty. A scheme, reported earlier, 6 

Fig. 4 Modified Eastman Type III camera showing syn- 
chronizing alternator. 

whereby a mechanical commutator was attached to the driving pulley 
of the Eastman camera, was abandoned in favor of a drag-free elec- 
tromagnetic pickup which operates on the variable-reluctance prin- 
ciple. A photograph of this pickup attached to the Eastman camera 
is shown in Fig. 4. Onto the face of the driving pulley of the Eastman- 
shutter system are placed nine, equispaced, thin iron wafers which 
rotate with the pulley. They can be seen in the photograph. Near by 
is mounted a small Alnico horseshoe magnet on which is wound a 
pickup coil. As the drive pulley rotates, the iron wafers pass in and 
out of the magnet's field, thereby causing reluctance variation in the 
magnetic circuit which results in a cyclic variation of the magnetic 

128 WHELAN March 

flux linking the coil wound on the horseshoe magnet. A voltage 
appears across the coil which is approximately sinusoidal in wave 
form, and whose instantaneous frequency corresponds to the in- 
staneous shutter frequency. The necessity for the use of nine iron 
wafers on the pulley arose because the drive pulley acts through a gear 
train to operate the shutter at just nine times its own rotary speed. 
This method of obtaining a synchronizing signal is sturdy and simple 
to adjust. It suffers from a sensitivity to stray 60-cycle pickup from 
the camera-driving motor which tends to phase-modulate the final 
synchronizing pulses. This problem was overcome by insertion of a 
60-cycle electrical filter in the output of this synchronizing system. 
If operation of the camera in the neighborhood of 60 frames per 
second is imperative, stray pickup balancing coils can be attached, 
and extreme shielding measures applied. 

Application of this method of synchronization to cameras other 
than the Eastman are under way. The actual technique must be 
varied from camera to camera in order to meet the specific mechani- 
cal requirements of each camera. It is essential that the point in 
the mechanical system where the signal takeoff is placed be gear- 
linked to the rotating-prism shutter so that relative motion between 
these two points is not possible. 

Before the low-level synchronizing signal from the camera can be 
utilized to trip the thyratron switch in the modulator, it must be 
shaped and amplified. At the same time, it is fed through dividing 
circuits which provide the interlacer flexibility mentioned previously. 
The divider circuit is a conventional "scale-of-two" counter which 
provides two outputs, each at one half the frequency of the input 
synchronizing signal. These two outputs are identical except that 
they are 180 degrees out of phase with each other, and consequently, 
while one of these outputs occurs on each even-numbered frame, the 
other occurs on each odd-numbered frame. Each of the divided out- 
put signals is exactly in phase with the wide-open position of the 
shutter, which assured uniform synchronization of all outputs, 
whether divided or not. 

Both the divided and undivided signal outputs are shaped into 
nearly rectangular pulses. The rise time of the pulse is 0.5 micro- 
second, or less, while the base width is 10 microseconds. Constant- 
amplitude synchronizing signals of nearly 100 volts is provided from 
each of the three output terminals of the divider-shaper while the 
camera is running. 


After division and shaping, the synchronizing pulses are delivered 
to the line amplifiers, which are six in number. Each of these amplifier 
stages consists of two cascade video units. The purpose of these line 
amplifiers is twofold: (a) to provide individual control points in the 
signal lines at which unblocking or interval control can be exercised, 
and (b) to provide electrically isolated trigger pulses at appropriate 
power levels for the thyratrons in the modulators. Care has been 
taken in the design of these line amplifiers to preserve the wave 
form of the synchronizing trigger pulses as they are delivered from the 
shaper-driver unit. Similarly, phase-delay errors have been mini- 
mized. Unblocking is provided in the line amplifier by means of elec- 
tronic switching of electrode voltages in the output tube of the in- 
dividual channels. 

The duration of the unblocked or operating cycle of each line ampli- 
fier is set by the use of stabilized multivibrator-type timing circuits 
which are adjustable in fixed time steps by the operator. 

A simulator unit is provided on the main control panel for "on-the- 
spot" investigations of trouble. Routine tests and specific symptoms 
have been outlined which permit relatively unskilled personnel to 
localize difficulties in the system. In general, it is the function of the 
simulator to generate artificial signals which resemble the actual 
operating signals, and these signals are then used as samples for 
sounding out the various portions of the system. This permits a 
step-by-step dynamic test of the various units. 

Another purpose the simulator serves is that of supplying a known 
number of flashes. These flashes constitute a standard test burst 
which permit the operator to check the intensity of his subject il- 
lumination by means of a suitable exposure meter of the integrating 
type. 6 


(1) Henry M. Lester, "Electronic flashtube illumination for specialized motion 
picture photography," /. Soc. Mot. Pict. Eng., vol. 50, pp. 208-233; March, 1948. 

(2) Robert T. Knapp, "Special cameras and flash lamps for high-speed under- 
water photography," J. Soc. Mot. Pict. Eng., vol. 49, pp. 64-82; July, 1947. 

(3) "Pulse Modulators," Book 9, MIT Radiation Laboratory Series, McGraw- 
Hill Publishing Company, New York, N. Y., 1948. 

(4) "Principles of Radar," MIT Radar School Staff, McGraw-Hill Publishing 
Company, New York, N. Y., 1946. 

(5) Robert A. Anderson and W. T. Whelan, "High-speed motion pictures with 
synchronized multiflash lighting," /. Soc. Mot. Pict. Eng., vol. 50, pp. 199-208; 
March, 1948. 

(6) An appropriate instrument has been developed and announced by H. E. 

Journal of the 

Society of Motion Picture Engineers 



Films in Television 363 

Possibilities of a Visible Music RALPH K. POTTER 384 

Optimum Performance of High-Brightness Carbon Arcs 


Effect of Carbon Cooling on High-Current Arcs 


Disk Recorder for Motion Picture Production J. L. PETTUS 417 

Synchronous Disk Recorder Drive C. C. DAVIS 427 

Test-Film Calibration Proposed Standards 


Proposed American Standard 447 

Your Society Report of the Executive Sacretary 


Officers of the Society 458 

Governors of the Society 460 

Officers and Managers of Sections 462 

Constitution and Bylaws 463 

Awards 474 

Society Announcement 478 

Report of the Treasurer 479 

Membership and Subscription Report 480 

Committees of the Society 481 

Section Meetings 495 

Book Review: 

"Friese-Greene : Close-up of an Inventor," by Ray Allister 

Reviewed by Terry Ramsaye 496 

Current Literature 498 


Chairman Editor Chairman 

Board of Editors Papers Committee 

Subscription to nonmembers, $10.00 per annum; to members, $6.25 per annum, included in 
their annual membership dues; single copies, $1.25. Order from the Society's General Office. 
A discount of ten per cent is allowed to accredited agencies on orders for subscriptions and 
single copies. Published monthly at Easton, Pa., by the Society of Motion Picture Engineers, 
Inc. Publication Office, 20th & Northampton Sts., Easton, Pa. General and Editorial Office, 
342 Madison Ave., New York 17, N. Y. Entered as second-class matter January 15, 1930, 
at the Post Office at Easton, Pa., under the Act of March 3, 1879. 

Copyright, 1949, by the Society of Motion Picture Engineers, Inc. Permission to republish 
material from the JOURNAL must be obtained in writing from the General Office of the Society. 
Copyright under International Copyright Convention and Pan-American Convention. The 
Society is not responsible for statements of authors or contributors. 

Society of 

Motion Picture Engineers 

342 MADISON AVENUE NEW YORK 17, N. Y. TEL. Mu 2-2185 

Earl I. Sponable 
460 W. 54 St. 
New York 19, N. Y. 

Peter Mole 

941 N. Sycamore Ave. 
Hollywood 38, Calif . 



Loren L. Ryder 
5451 Marathon St. 
Hollywood 38, Calif. 

Clyde R. Keith 
120 Broadway 
New York 5, N. Y. 



VICE-PRESIDENT R a i p h B. Austrian 

John A. Maurer 1270 Avenue of The 
37-0131 St. Americas 

Long Island City 1, N. Y. New York 20, N. Y. 

Robert M. Corbin 
343 State St. 
Rochester 4, N. Y. 

William C. Kunzmann 
Box 6087 
Cleveland 1, Ohio 


David B. Joy 
30 E. 42 St. 
New York 17, N. Y. 


Alan W. Cook 
25 Thorpe St. 
Binghamton, N. Y. 

Lloyd T. Goldsmith 
Warner Brothers 
Burbank, Calif. 

Paul J. Larsen 
508 S. Tulane St. 
Albuquerque, N. M. 

Gordon E. Sawyer 
857 N. Martel Ave. 
Hollywood 46, Calif. 


James Frank, Jr. 
426 Luckie St., N. W. 
Atlanta, Ga. 

William B. Lodge 
485 Madison Ave. 
New York 22, N. Y. 

William H. Rivers 
342 Madison Ave. 
New York 17, N. Y. 

Sidney P. Solow 
959 Seward St. 
Hollywood 38, Calif. 

R. T. Van Niman 
4431 W. Lake St. 
Chicago 24, 111. 


Herbert Barnett 
Manville Lane 
Pleasantville, N. Y. 

Fred T. Bowditch 
Box 6087 
Cleveland 1, Ohio 

Kenneth F. Morgan 
6601 Romaine St. 
Los Angeles 38, Calif. 

Norwood L. Simmons 
6706 Santa Monica Blvd. 
Hollywood 38, Calif . 

Films in Television 



Cameras At the present time films for television are being photo- 
graphed with both 35-mm and 16-mm motion picture cameras at the 
standard speed of 24 frames per second. For production work where 
synchronized sound is to be used, the camera must be driven at syn- 
chronous speed. A number of television stations currently making 
their own newsreels use commercially available 16-mm professional 
cameras and associated equipment. 

Composition Data supplied by one television station indicated 
that because adjustment of picture size in home receivers varies 
greatly, all significant action and subject material be kept within a 
central area having 8 l /% per cent top and bottom margins and 13 per 
cent side margins. When this is done, a large majority of commercial 
receivers will show all-important information. 

Close-up scenes give most pleasing reproduction because viewing 
screens of home receivers are small and the field of action necessarily 
is limited. Medium shots are generally considered the outside limit 
and long shots rarely add anything of value to the film program. 
Subject matter should be kept as large as the limits and action of the 
scene being televised will allow without obvious crowding of action or 

Whenever possible checkerboard patterns with many abrupt 
changes of contrast should be employed as these numerous large vari- 
ations in print density will reduce the horizontal-smear effect that 
otherwise would be caused by low-frequency defects of present sys- 
tems. For the same reason, large uniform-colored or relatively dark 
areas and delicate or minute patterns are to be studiously avoided, 
particularly in the lower portion or foreground of the scene. 

Subject Lighting The limited range of picture-tube brightness re- 
quires that subject contrast be controlled wherever possible. Usually 
it is not necessary to resort to flat lighting in order to. hold contrast 
within 'the brightness range of the television system, but even lighting 


is essential particularly over large picture areas. That is, large pic- 
ture areas must have about the same average illumination. Wide 
variations in brightness over the scene will otherwise have to be com- 
pensated for by adjustment of the television shading controls. 

Adequate foreground lighting is quite essential since the electric- 
energy-decay-rate characteristic of the iconoscope mosaic may cause 
picture degradation in the form of insufficient signal response in the 
lower portion of the received picture. The general intensity of illumi- 
nation from scene to scene should be kept relatively constant so that 
the level of the television signal does not change markedly and for this 
reason night scenes should be avoided. For psychological reasons 
long fades should not be used because they interrupt program con- 
tinuity and the audience may think from the long blank period that 
something is wrong with the receiver. 

Properties Clothing and accessories, backgrounds, furniture, and 
other "properties" should have definite patterns large enough to be 
clearly visible on the screen of the television receiver. Again, fine or 
delicate detail with minute changes in contrast should be avoided. 

Titles To reproduce clearly on small home receivers, the lettering 
of titles should be large, boldface on a textured background, and 
should always be located within the dimensional limits previously 

General Action within scenes should be continuous. This, how- 
ever, is not always possible, so where inanimate objects are shown for 
any period of time, motion of the camera by zooming, traveling, 
change of angle, or slow panning should be substituted to accomplish 
the desired effect. In the present state of the art, this type of change 
sometimes emphasizes the geometrical distortion in the final image. 
In the transition from one scene to the next, it is desirable to employ 
lap dissolves, quick fades, or instantaneous "cuts" timed to keep pace 
with the program. 


j S5-Mm Negative Normal exposure and development, as employed 
in motion picture negative work, should be used for pictures to be 
televised. Negative gamma is usually carried between 0.65 and 0.70 
and the scene density is considered normal if the negative prints in the 
middle of the printer scale. 

85-Mm Prints Over a period of years numerous closed-circuit 
tests have been run in an attempt to determine optimum print density 


for televising. These tests, although they were not conclusive, have 
shown that low-contrast prints (gamma between 1.4 and 1.6) with a 
general density near normal reproduce well. When the contrast was 
carried to normal (gamma of 2.20 to 2.50) the print which reproduced 
best was at least two printer points light. 

All of these tests were made from a negative exposed for motion 
picture theater use. More recent tests have shown that prints of 
normal gamma and perhaps 1 to 2 printer points light reproduce 
best. In view of the great importance of establishing proper film 
specifications for television this subject needs further investigating 
and reporting. 

16-Mm Reversal Most 16-mm film used by television stations is 
processed by reversal. Current 35-mm practice shows that a nega- 
tive gamma of 0.70 and a print gamma of 1.50 produce a resulting 
picture contrast of 1.05, while current 16-mm reversal technique pro- 
duces a print gamma between 1.00 and 1.20, which has proved satis- 
factory and is recommended. 

16-Mm Negative and Positive A limited amount of 16-mm negative 
and print work is being done. Current practice is to develop the 
negative in fine-grain negative developer and print normally. 


The translation of motion pictures into television signals is compli- 
cated by the fact that motion picture film moves at the standard rate 
of 24 frames per second while the rate of the television signal is 30 
frames (60 fields) per second. A simple factor can be applied to the 
different frame rates which satisfies the peculiar characteristics of the 
two systems. Two frames of motion picture film require the same 
amount of time as five fields (2 J /2 frames) in television scanning. This 
relationship is presented graphically in Figs. 1A and IB, which show 
that if one film frame is scanned for two television fields and the next 
film frame for three television fields the time difference of frame rate 
can be satisfied. This relationship is fundamental as long as the re- 
spective frame rates are retained and applies regardless of type of 
camera or projector. 

There are two fundamentally different types of television pickup 
tubes, the storage type (iconoscope image orthicon) which stores 
electrical charges produced by a multitude of individual picture ele- 
ments until discharged by the scanning electron beam; and the non- 
storage type (image dissector phototube) where the electrical energy 




I/ 24 SEC- 0.0417 SEC 



-* 0.0083 K- 0-0125 0.0083 

A. 35-mm motion picture projector with 72-degree shutter, 24 
frames per second. 

1/30 SEC = 0.0333 SEC 

1/60*0.0166 I/60-O.OI66 


1 III 


B. Television picture with interlace scanning, 30 frames per second. 


'/30 SEC 


, LIGHT v 

1 / PULSE \ 

LLDOWN ./ \jl n 

m^M | A 




C. 35-mm television 1 intermittent storage system of scanning 
motion picture film. 

'/30 SEC 

1/30 SEC 



D. Television scanning of motion picture film run at 30 frames per 
second, 72-degrees pulldown time. 

Fig. 1 


of each picture element is proportional to the incident light experi- 
enced at the instant that element is scanned. The phototube is used 
with the flying-spot scanner and is gaining in popularity with de- 
velopment engineers. 

The iconoscope camera tube, however, is almost universally used 
for commercial-film pickup work. Because of its storage feature the 
iconoscope can be "pulsed' ' with an intense burst of light of short 
duration. This produces a charge picture in the tube that is then re- 
moved in the normal scanning sequence. This flash may not be ap- 
plied during the actual scanning time since it would give a pulsed 
video signal and a noticeable black "application bar" across the re- 
ceiver screen. Light is therefore applied during the vertical-blanking 
period and its pulse effect is further nullified by proper back lighting 
of the mosaic screen in the iconoscope tube and electronic gating of 
the beam current. Since light is applied only during vertical blanking 
a full scanning interval is available for pulldown of the next film 
frame. Fig. 1C indicates the sequence of charging the camera tube 
with a light pulse, scanning the resulting picture, and film pulldown in 
35-mm projectors. Either mechanical or electrical means can provide 
the pulse. A pulldown of approximately 50 degrees and a mechanical 
shutter having an opening of less than 18 degrees and synchronized 
at 3600 revolutions per minute to open during the television vertical- 
blanking pulse time is practical for 16-mm projectors. Equipment is 
also available with an electrically timed and controlled gas-discharge 
tube instead of a mechanical shutter. 

Control of the iconoscope camera requires adjustment of the beam 
current and continual monitoring of picture "shading." Beam cur- 
rent can be set for average light level, compromising between exces- 
sive tube noise at high beam levels and low signal with resulting 
amplifier noise at low beam levels. Shading, an undesirable charac- 
teristic, is a spurious signal resulting from an uneven distribution of 
secondary electrons on the tube mosaic and varies with picture con- 
tent. Adequate correction can be obtained by properly mixing arti- 
ficially generated signals, saw-tooth and parabolic, and occasionally 
some sine-wave forms in both the vertical- and horizontal-scanning 
directions, and applying the results to the camera output. Another 
difficulty known as edge flare, which shows up as bright areas usually 
on the right edge and bottom of the picture, can be improved by the 
adjustment of an internal edge lighter. 

With proper adjustment of the controls and proper high-light 


illumination of the iconoscope mosaic a very satisfactory picture is 
obtained. Resolution usually exceeds 350 television lines and the sig- 
nal-to-noise ratio is low but tone gradation is not perfectly linear. The 
signal-output current is approximately proportional to brightness of 
mosaic illumination up to about 0. 1 foot-candle, but at brighter levels 
the signal increases less rapidly with increasing brightness. Thus in 
combination with the normal viewing tube both the blacks and the 
whites seem to be compressed. 

The coating on the mosaic of the Type 1850-A iconoscope shows a 
preference for the blue region of the spectrum so that color films can 
be projected for black-and-white television pickup but tone values of 
various colors will not agree perfectly with those seen by the eye. 
Some partial correction is possible by the use of filters on the light 

The phototube flying-spot-scanner system is now undergoing de- 
velopment and shows considerable promise. It has a number of very 
desirable advantages over the iconoscope for film pickup, namely, 
simplicity of components, freedom from shading and other spurious 
signals, no loss of stored charge during the scanning cycle, excellent 
contrast range, and high picture resolution. 

A major difficulty of the flying-spot scanner, as in any type of non- 
storage television camera tube, as shown by Fig. IB, is that no film 
pulldown time is available when the projector is run at 30 frames per 
second. Some type of nonintermittent projector would seem to be 
desirable but the complexity as well as unsatisfactory speed regulation 
of several proposed types of continuous projectors presents a serious 

The iconoscope camera and the flying-spot scanner are both useful 
with still slides or filmstrips in a standard projector. Camera switch- 
ing can be accomplished by remote control and in one case, two film 
projectors and a slide projector can be switched into a single camera 
by the use of an accessory optical-mirror device. Television lends 
itself nicely to "fades," "dissolves," and superposition of two pictures 
by the simple expedient of mixing video signals at the required level 
before adding the standard synchronizing signal. Necessary controls 
are commercially available as standard studio equipment. "Wipes" 
are somewhat more difficult, requiring an electronic switch of a type 
that is not as yet commercially available. 

Sound for television from film sources requires no special handling 
beyond equalization. 



Motion pictures photographed from a television picture tube are 
made as transcriptions of live-studio or remote programs for rebroad- 
casting and may be used at a later time by the station that presented 
the original program or may be syndicated with several prints from 
the original made for distribution to subscriber stations. Picture- 
and sound-quality requirements are high, demanding utmost atten- 
tion on the part of station and processing laboratory personnel. 

Regular record films are also made but generally at a reduced film- 
frame rate and have far less rigid quality requirements because they 
are never rebroadcast. 


The conversion from the 30-frame-per-second television-picture 
rate to the 24-frame-per-second film-picture rate presents a serious 
problem for television recording-camera design engineers. A cur- 
rently successful solution is based on the use of successive dissimilar 
scanning cycles. Another proposed answer is a change of the stand- 
ard film rate from 24 to 30 frames per second. The logic of this solu- 
tion appears obvious but there is a serious handicap of economic 
inertia to consider since sound films have been made at 24 frames per 
second and studios and theaters have been following the present stand- 
ard for over 20 years. There is also the problem of providing pull- 
down time if photography is on an intermittent basis. If a film rate 
of 30 frames per second is ever adopted, it appears that some method 
of continuous film motion will be desirable, if the necessary constancy 
of motion can be obtained. 

It is possible to design cameras that use either mechanical or elec- 
trical blanking during the pulldown period. Continuously moving- 
film cameras are also possible but the mechanical, optical, and syn- 
chronization problems involved are most difficult. 

One 16-mm television-recording camera now in use is equipped with 
a mechanical shutter driven by a synchronous motor from the same 
60-cycle alternating-current power source as is used for the television- 
synchronizing generator. This shutter has a closed angle of 72 de- 
grees and an open angle of 288 degrees. At the 24-cycle rate this 
represents a closed time of Vi2o second and an open time of Vao second. 
The latter is equivalent to one full television-frame cycle. 




Fig. 2 shows the time sequence of this shutter in relation to the 30- 
frame (60-field) television-scanning cycle. The camera shutter re- 
mains open for exactly two television fields, closes for exactly J /2 field 
while the film is advanced, then opens again for the exact equivalent 
of two more television fields (actually */2 plus 1 full plus */2 field). It 
then closes for l / 2 field while the film is advanced a second time and 
again opens at exactly the beginning of the next field. The two non- 
symmetrical cycles are then repeated. 

One serious objection to the mechanical shutter for television-picture 
recording lies in the need for perfect synchronization between the 
motor that drives the shutter and the television frame-rate generator 






* i/30 SEC. *\ 

^-j/boscc. M 

I 1 1 



U . 'I' 2 

J 4 J5 6 


1 I 
*- -H DOWN ** 




2 Time sequence of exposure and pulldown timing of the camera in 
relation to the field rate of the television image. 

which may not necessarily operate from the same 60-cycle alternating- 
current power line. The shutter action is critical in that it must 
rotate with extremely low flutter content since minute changes in 
angular velocity will result in banding, the effect of over- or under- 
exposure of scanning lines adjacent to the cutoff point. 

With the electronic shutter now being used with some installations, 
this problem is minimized because the television-picture tube is elec- 
tronically blanked or turned off at the end of each 525 lines (one com- 
plete television-frame cycle) and is not turned on again until the film 
has been pulled down and brought to rest. Also, the electronic shut- 
ter can accommodate any film-frame rate less than a given maximum 
determined by the practical limitations of film-pulldown time. 



The majority of television-film recordings are made on 16-mm 
rather than 35-mm film. The major reason is economic, since the 
cost of 35-mm film is somewhat more than three times the cost of 16- 
mm per unit of recording time. The current quality of television 
images, which undoubtedly will undergo gradual refinement, is con- 
sidered to be roughly equivalent to 16-mm home motion pictures. 
No marked improvement, however, is to be had by recording on 
35-mm rather than 16-mm film at the present time. With the use of 
fine-grain, high-resolution, 16-mm-film emulsions, no loss of resolution 
in recording the television image is noticeable. 

Fire regulations covering the use of 35-mm film, which apply re- 
gardless of whether the 35-mm film is acetate safety base or the com- 
bustible nitrate base, are rigorous. The cost of providing space that 
meets these regulations for the use of 35-mm film is extremely high 
and the changes needed in existing space are difficult to accomplish. 
Sixteen-millimeter films are available only in acetate safety base 
which is classified by the Underwriters' Laboratories as having a 
safety factor slightly higher than that of newsprint. The use of 
16-mm films, therefore, is not restricted by fire regulations. It 
should be noted that in New York City these restrictions apply to 
space in which equipment capable of operating with 35-mm film is 
installed, so in order to forestall trouble, all equipment should be 
single-purpose, 16-mm equipment rather than dual-purpose, 35-mm 
or 16-mm equipment. 

Another factor in the choice of 16-mm film is the high cost of 35-mm 
projection equipment. Most television stations are providing pro- 
jection facilities for 16-mm film only for this reason. In order to 
service these stations with syndicated programs photographed from 
the picture tube, 16-mm prints will be needed. 


Film Emulsion There are three general classifications of film 
emulsions in terms of their spectral characteristics and they can be 
matched to the phosphor spectral characteristic of the television-pic- 
ture tube, for greatest actinic efficiency. 

1. Panchromatic emulsions are most sensitive in the range from 
the ultraviolet (4000 angstrom units) through the red (7000 angstrom 


units) . The spectral response of these emulsions corresponds approxi- 
mately to that of the eye and so they are generally used for direct 

2. Orthochromatic emulsions are sensitive from the ultraviole- 
through green (5700 angstrom units) and are used in direct photograt 
phy where it is desirable to reduce the red sensitivity ; 

3. " Ordinary," blue-sensitive emulsions, respond to the ultraviolet 
and blue portions of the light spectrum. This type of emulsion is used 
in coating films and papers generally employed in making positive 
prints from negatives. It is economical in comparison to the panchro- 
matic and orthochromatic types. Another advantage is the ease of 
handling as relatively bright safelights may be used. 

Picture-Tube Phosphors To match these film characteristics, pic- 
ture-tube phosphors are available with light output ranging from 
the ultraviolet through the entire visual spectrum. Three types of 
phosphors in common use in television techniques are as follows: 

1. PI, green fluorescence, commonly used in oscillographic work. 
It is the most efficient visually, but has poor actinic efficiency. 

2. P4, white fluorescence, used for black-and-white reproduction 
of television images in most home receivers. It has the advantage in 
picture-tube photography that picture quality is most readily judged 
visually. However, some P4 screens have undesirable decay charac- 

3. P5 and PI 1 ; these two phosphors are blue with high ultraviolet 
output. Photographically, they are very efficient. There is the 
difficulty in using a blue phosphor in judging the quality of image 
visually, because of the fact that the human eye has a low response in 
the blue region and cannot evaluate the quality of the. ultra violet 
component of the image-light output at all. 

Tests have indicated that for recording of television images a blue- 
fluorescing screen (P5 or Pll) is desirable since it makes possible the 
use of high-resolution, low-cost, positive types of film stocks. The 
P5 screen has excellent persistence characteristics but produces a 
somewhat lower light level than that which can be obtained with Pll. 


A method for establishing brightness range and exposure level is as 
follows : A plain raster is used on the tube such as would be obtained 
by the use of the blanking signal or pedestal without picture modula- 
tion, The brightness of this raster is varied by means of the video 


gain control or picture-tube, grid-bias control. The beam current is 
measured by means of a microammeter. Since the light output of the 
tube is dependent upon the power input to the screen, the measure of 
beam current affords a measure of the brightness of the tube. Film is 
exposed to this raster with the beam curren ts varied in steps . The den- 
sity of the film processed as a normal negative is measured and plotted 
against the logarithm of the beam current. A normal negative de- 
veloped to a gamma of 0.65, which has been exposed to an object with 
a brightness range of 1 to 30 (in logarithmic increments, a range of 
1.5) should have a density range from 0.25 in the shadows to approxi- 
mately 1.4 in the high lights. The change in beam current necessary 
to produce such a range on the picture tube can be read from the plot 
of the log of the beam current and film density. The average bright- 
ness of the cathode-ray tube with picture then would be set by using a 
beam current that produces a density in the middle of the above 
range. The video signal is adjusted to a level that will put the blank- 
ing level of the composite signal just at visual cutoff of the cathode- 
ray tube. A picture signal judged to have an alternating-current 
axis of 50 per cent should be used for this adjustment. This method 
is largely empirical, but, with experience on the part of the operator, 
can be made to give consistent results. 


A number of tests have been made in co-operation with the film 
manufacturers on the processing and printing of films photographed 
from a television-picture tube. Both reversal and negative processing 
of the original film were tried and results show that standard process- 
ing methods result in optimum picture quality. Negatives exposed 
to television images originating in iconoscope cameras are developed 
to a gamma of 0.7 as determined by a standard lib sensitometric test. 
Film of orthicon pickups gives best results when processed to approxi- 
mately 0.6 or 0.65. These are interim values as tests on the process- 
ing of these films have not been completed. 

Printing is done according to standard motion picture laboratory 
practice. Step printing in which the print stock and negative are ex- 
posed to the printing light a frame at a time is preferred over continu- 
ous printing, where the negative and print stock run past an illumi- 
nated slit at a continuous speed. There is a sufficient amount of slip- 
page between the negative and the print stock in the continuous 
printing process to degrade the resolution of the television image. 


Contrary to the opinion held by many workers, the fact that the film 
picture of a television image is poorer in resolution than in the case of 
direct photography does not mean that less care is required in the 
handling of the film in printing and in projection. The fact is that 
the utmost care must be taken to maintain the original quality inher- 
ent in the film negative throughout the printing process and in the 
projection of the resulting print. 

Films of iconoscope programs usually can be printed at one printer- 
light setting, that is, the densities and contrast range of the film re- 
sulting from the recording of the outputs of a number of iconoscope 
cameras do not change sufficiently to warrant changes in the intensity 
of the printing light. 

In film recordings of programs picked up by orthicon cameras the 
picture negative often has to be timed for printing. Frequently there 
is some difference in the brightness range between different orthicon 
cameras. Much of this change can be charged to the fact that the 
spectral characteristics of the orthicon may vary from tube to tube. 
An orthicon with high infrared response has a somewhat different 
tonal graduation than an orthicon with lower response in this region. 

In recording for retransmission through the television system a 
print gamma of 2.2 and a maximum density of 2.4 have been found 
satisfactory. Further tests may show the desirability of changing 
these recommendations, but to date the best results in the televising 
of release prints have been obtained under such conditions. 

Emulsion position in the final print is of importance in television 
because films may be spliced with other films for special purposes. 
The use of a nonstandard emulsion position requires a change of focus 
in the film projector when interspaced with films using a standard 
emulsion position. This would require the constant attention of the 
projectionist to maintain optimum focus throughout the spliced film; 
therefore it is advantageous to insist upon a standard emulsion posi- 
tion for all film to be used in television. The American Standard for 
16-mm film is emulsion "toward the screen." 

In, the recording of television images there are several methods of 
obtaining the final print : 

1. The use of reversible film stock in photographing a positive 
cathode-ray-tube image. A dupe negative may be made of this 
material from which additional prints can be made. The final prints 
then have standard emulsion position; 


2. Photography using high-contrast positive stock and a negative 
picture-tube image resulting in a positive print from which dupe 
negatives may be made if production prints are required. These 
prints will have standard emulsion position ; 

3. The use of a positive image, photographing with a negative 
type of film from which final prints are made, resulting in a nonstand- 
ard emulsion position. (By reversing the direction of horizontal scan- 
ning, however, the original negative may be made to have the same 
emulsion position as that of a dupe negative. Prints made from this 
negative then have standard emulsion position.) 

When pro'duction prints are required Method 3 is now used almost 
exclusively since it eliminates the use of a dupe negative and conse- 
quently introduces less total degradation. Methods 1 and 2 do not 
produce production prints of suitable quality for present-day commer- 
cial television. 


An early system of film scanning was described by Ives (1931) as 
incidental to a three-channel system of television. The three chan- 
nels were used to obtain the desired resolution, without increasing the 
frequency bandwidth beyond the technique of the art then available 
(40 kilocycles). The three channels were optically separated into 
three independent interlaced fields, giving 108 lines in all. The film 
was standard 35-mm, drawn continuously past a mechanically scan- 
ning Nipkow disk at 18 frames per second. No mention is made of 
how this was reconciled to the standard frame speed. 

A later film scanner, also attributed to Ives (1938), was used to test 
the Bell System coaxial cable from New York to Philadelphia for tele- 
vision transmission. This used 240 sequentially scanned lines at the 
standard 24 frames per second. It also employed standard 35-mm 
film, drawn continuously past a Nipkow disk. The disk, however, 
was fitted with lenses instead of holes. Here there was no problem of 
frame-rate conversion. 

An elaborate development was carried out in Germany by Fernseh 
AG (1939) on an intermediate-film quick processing device adapted 
to be used both for pickup of news and similar events, and for theater 

The pickup device used standard 35-mm film, but the exposed frame 
was hah 7 size in both dimensions, to save film; 16-mm film was not 
used because it lacked the strength necessary for the quick processing 
baths. In its later form it followed the German 441-line, interlaced- 


scanning, 25-frame-per-second, television standards. This used a 
multiple-spiral Nipkow disk with continuously moving film, with no 
frame-rate conversion problem (the film being taken at 25 instead of 
24 frames per second). It was in a later form replaced by a dissector 

An extensive photographic investigation was made for the quick 
processing. A special thin emulsion was used. In the later model the 
processing times were 

Development 5.0 seconds 

Intermediate bath 2.5 seconds 

Fixing 15.0 seconds 

Washing 10.0 seconds 

Drying 43.0 seconds 

The negative film was scanned directly after drying. 

In one unusually elaborate form of the apparatus the film, after 
using, was again washed, scraped free of emulsion, dried, coated with 
fresh emulsion, dried, and used again. 

The whole equipment was set up in a special television truck, a 
series of which was built. 

The intermediate-film projector was designed for the earlier 180- 
sequential-line scanning at 25 frames per second. It used split film, 
17.5 mm wide, with an 8- X 11-mm image. This was a positive, 
taken from a negative image on a 12-kilovolt cathode-ray tube. The 
camera used intermittent film motion synchronized with the tele- 
vision. The processing times were 

Development 24 seconds 

Fixing 24 seconds 

Washing 12 seconds 

Drying (not stated) 

The projection was on to a 2.2- X 3-meter screen. Because of the 
high film cost and the rapid advances in projection tubes the German 
intermediate-film projector was abandoned. 

A film scanner attributed to Jensen (1941) was used by the Bell Sys- 
tem for testing the prewar television transmission circuits over coaxial 
cable. This used standard size but specially printed 35-mm film. 
The film was drawn continuously past a gate and focused on the 
photosensitive cathode of an image-dissector pickup tube. Extensive 
study was made of the focusing and deflecting coils in the latter, to 
obtain improved results. The special printing of the film was used to 




obtain the frame-speed conversion and interlacing required, with con- 
tinuous film motion and mechanical simplicity. In the specially 
printed film one frame is used for each television field scanning. Thus 
it is obtained from the original film by printing its odd-numbered 
frames twice in succession, and its even-numbered frames three times 
in succession; two successive frames of the original thus occupying 
five frames in the print. This enabled the television signal to follow 
the then current 441-iriterlaced-line, 30-frame-per-second standards. 
The blanking period was adjusted from the film standard to the tele- 
vision standard by a slight compensating vertical sweep in the image- 
dissector tube. The light source used was a 1000-watt incandescent 

The sound track was printed specially on the film also, ' "stretched" 
in the direction of motion in the ratio of 2.5 to 1. Aside from this, 
the sound pickup was standard. 


An early system of color-film scanning is also attributed to Ives 
(1931). This made use of 16-mm Kodacolor film of that time. Film 
motion was continuous and the scanning mechanical with a Nipkow 
disk. The colors were led separately by lenses and mirrors to three 
phototubes (so that the system was of the simultaneous type). Be- 
cause of the nature of the Kodacolor film, with its lenticular markings 
on the film surface, the color separation was already obtained geo- 
metrically, and no filters were necessary. The television standards 
were 50 lines, 18 frames per second, which previously had been used 
with black-and-white. The received signal was reproduced on three 
lamps, superposed optically on a Nipkow scanning disk, and viewed 
monocularly through an eyepiece by a single observer. 

The Columbia Broadcasting System for some years has been in- 
tensively developing a color system. On September 27, 1946, this 
was proposed in a petition to the Federal Communications Commis- 
sion as the basis for a commercial broadcast service in color television. 
After an extensive hearing on the subject, however, the petition was 
denied on March 18, 1947. 

The Columbia system, in so far as it used film, has principally used 
16-mm film because it was expected that the major available material 
would be in that size, but a 35-mm machine using the same principles 
has also 'been in preparation. The film in each case is standard and 
operates at 24 frames per second. The film is driven continuously 


past a special optical system using an arc lamp which focuses 
successive fields, as they are to be scanned, on to the photosensitive 
cathode of an image-dissector tube. These act in co-operation with a 
suitable compensating vertical sweep with this tube, to give the cor- 
rect interlaced scan required. The CBS proposed standard calls for 
144 fields (or scans) so that a film-frame to scanning-field conversion 
of 24 to 144 (or 1 to 6) is required. This is accomplished with using a 
special optical system which allows each film frame to be scanned six 
times as the frame moves past the gate aperture. There is a special 
adjustment for differing film shrinkage, which involves a change of 
magnification and a refocusing of the photosensitive cathode. 

The CBS system is arranged with a set of six fixed-color filters 
through each of which the beam is directed in turn by the special op- 
tical system. These function in the same manner as would a rotating 
tricolor disk and allow successive fields to be scanned in the successive 
three color primaries. Six field scannings are necessary. The signal 
transmits the picture fields in the successive colors sequentially. 
However, a synchronous arrangement is also provided for adjusting 
the signal gain for any one color independently, to permit modifying 
the color balance while the apparatus is running. 

The sound pickup is conventional. 

Principally to obtain certain elements of flexibility not easily per- 
mitted by .a sequential color system, the Radio Corporation of Amer- 
ica has worked on a simultaneous color system, and has demonstrated 
experimental versions of it on various occasions. 

In this case, 16-mm color film is used, which is driven past a flying- 
spot cathode-ray-tube scanner. The beam, after passing through the 
film, is separated by special mirrors via three filters to three non- 
storage phototubes, each of which generates one of the three simul- 
taneous signals to be propagated. 

The standards for each color have been taken by RCA to be the 
same as present black-and-white broadcast standards, namely, two 
interlaced fields per frame, at a frame rate of 30 per second. In fact 
the green channel is arranged to be used to reproduce a black-and- 
white picture, this being one of the items of flexibility desired. In the 
experimental demonstrations which have been given there has been 
no provision for frame-rate conversion, so that the action in the film 
is speeded up in the ratio of 30:24. Similarly no arrangement has 
been provided for adjusting the television to the film-blanking period 
(which latter is nearly zero in 16-mm film) so that a portion of the 
frame appears black in the reproduction. 





This report, prepared by the Television Committee of the So- 
ciety of Motion Picture Engineers, contains information on points 
of common interest to television and motion picture engineers. 

Because of the rapidly changing state of the art, it was found im- 
possible to present complete data on all aspects of films in television. 
An attempt has been made, however, to present as much information 
as possible concerning present practices in the hope that it will serve 
as a guide and aid to those seeking information in this field. 

The membership of the committee is as follows : 

D. R. WHITE, Chairman 
Du Pont 


Television Consultant 

National Carbon Company 

Warner Brothers Pictures 



General Electric Company 

Hazeltine Electronics Corporation 

Columbia Broadcasting System 

General Precision Laboratories 

International Projector Corporation 


Columbia Broadcasting System 


Allen B. DuMont Laboratories 

International Projector Corporation 

Paramount Pictures 

RKO Theaters 


Loew's Theaters 


Farnsworth Research Corporation 

Warner Brothers Pictures 


Columbia Pictures 

Don Lee Broadcasting System 

Bell Telephone Laboratories 


Allen B. DuMont Laboratories 

Metro-Goldwyn-Mayer Studios 

Bamberger Broadcasting Service 

Paramount Pictures 

Eastman Kodak Company 

Samuel Goldwyn Studio Corporation 

National Broadcasting Company 


Eastman Kodak Company 



The following is a bibliography of material dealing with the relationships be- 
tween television and motion pictures. It has been divided into the following 

1. General 3. Film from Television Sources 

2. Television from Film Sources 4. Color Television 

Under "general" are classified not only discussions dealing with the broader 
subject, but also those covering more than one of the subsequent topics. 

This bibliography has been quickly gathered from such sources as have been 
readily available, and does not purport to be complete. 


(1) Alfred N. Goldsmith, "Theater television A general analysis," J. Soc. 
Mot. Pict. Eng., vol. 50, pp. 95-122; February, 1948. 

(2) Ralph B. Austrian, "Showmanship side of television," J. Soc. Mot. Pict. 
Eng., vol. 49, pp. 395-405; November, 1947. 

(3) W. V. Wolfe, "Report of the SMPE Committee on Progress," /. Soc. Mot. 
Pict. Eng., vol. 48, pp. 304-317; April, 1947. 

(4) "Statement of SMPE on revised frequency allocations," /. Soc. Mot. Pict. 
Eng., vol. 48, pp. 183-203; March, 1947. 

(5) Lester B. Isaac, "Television and the motion picture theater," J. Soc. Mot. 
Pict. Eng., vol. 47, pp. 482-487; December, 1946. 

(6) Albert Rose, "A unified approach to the performance of photographic 
film, television pickup tubes, and the human eye," /. Soc. Mot. Pict. Eng., vol. 47, 
pp. 273-295; October, 1946. 

(7) Allen B. DuMont, "The relation of television to motion pictures," /. Soc. 
Mot. Pict. Eng., vol. 47, pp. 238-248; September, 1946. 

(8) P. J. Larsen, "Report of the Committee on Television Projection Practice," 
J. Soc. Mot. Pict. Eng., vol. 47, pp. 118-120; August, 1946. 

(9) Judy Dupuy, "Television Show Business" (book review), J. Soc. Mot. 
Pict. Eng., vol. 46, p. 424; May, 1946. 

(10) A. Rose, "Photographic film, television pickup tubes, and the eye," 
Intern. Proj., May, 1946. 

(11) "Technical News," /. Soc. Mot. Pict. Eng., vol. 46, p. 81, January, 1946. 

(12) Ralph B. Austrian, "Film The backbone of television programming," 
/. Soc. Mot. Pict. Eng., vol. 45, pp. 401-^14; December, 1945. 

(13) Ralph B. Austrian, "Some economic aspects of theater television," J. Soc. 
Mot. Pict. Eng., vol. 44, pp. 377-386; May, 1945. 

(14) Paul J. Larsen, "Statement presented before the Federal Communications 
Commission relating to television broadcasting," /. Soc. Mot. Pict. Eng., vol. 44, 
pp. 123-128; February, 1945. 

(15) "Technical News," /. Soc. Mot. Pict. Eng., vol. 43, pp. 303-304; October, 

(16) Worthington C. Miner, "Film in television: Television production as 
viewed by a radio broadcaster," /. Soc. Mot. Pict. Eng., vol. 43, pp. 79-93; August, 


(17) Wyllis Cooper, "Film in television: Television production as viewed by a 
motion picture producer," /. Soc. Mot. Pict. Eng., vol. 43, pp. 73-79; August, 

(18) "Television report, order, rules, and regulations of the Federal Communi- 
cations Commission," J. Soc. Mot. Pict. Eng., vol. 37, pp. 87-98; July, 1941. 

(19) "Report of the Television Committee" (Flicker, Visual Fatigue, Bibli- 
ography), /. Soc. Mot. Pict. Eng., vol. 35, pp. 569-584; December, 1940. 

(20) M. W. Baldwin, "The subjective sharpness of simulated television 
images," Bell Sys. Tech. Jour., vol. 19, p. 563; October, 1940. 

(21) P. C. Goldmark and J. N. Dyer, "Quality in television pictures," /. Soc. 
Mot. Pict. Eng., vol. 35, pp. 234-254; September, 1940. 

(22) A. M. Skellett, "Transmission system of narrow band- width for animated 
line images," J. Soc. Mot. Pict. Eng., vol. 33, pp. 670-677; December, 1939. 

(23) G. Schubert, W. Dillenburger, and H. Zschau, "Das Zwischen Film 
verfahren," Fernseh A. G. Hausmitteilungen, vol. 1, Part I, p. 65; April, 1939: 
Part II, p. 162, August, 1939; Part III, p. 201, December, 1939. 

(24) R. Holier and G. Schubert, "Zehn Jahre Fernsehtechnik," Fernseh A. G. 
Hausmitteilungen, vol. 1, p. Ill; July, 1939. 

(25) "Report of the Television Committee," J. Soc. Mot. Pict. Eng., vol. 33, 
pp. 75-80; July, 1939. 

(26) G. L. Beers, E. W. Engstrom, and I. G. Maloff, "Some television problems 
from the motion picture standpoint," J. Soc. Mot. Pict. Eng., vol. 32, pp 121-139; 
February, 1939. 

(27) A. D. Blumlein, C. 0. Browne, N. E. Davis, and E. Green, "The Marconi- 
EMI television system," J.I.E.E. (London), p. 758; December, 1938. 

(28) Herbert E. Ives, "Transmission of motion pictures over a coaxial cable," 
/. Soc. Mot. Pict. Eng., vol. 31, pp. 256-273; September, 1938. 

(29) M. E. Strieby, "Coaxial-cable system for television transmission," Bell 
Sys. Tech. Jour., vol. 17, p. 438; July, 1938. 

(30) "Television demonstration at the fall convention," /. Soc. Mot. Pict. Eng., 
vol. 29, pp. 596-603; December, 1937. 

(31) "Television from the standpoint of the motion picture producing indus- 
try," J. Soc. Mot. Pict. Eng., vol. 29, pp. 144-149; August, 1937. 

(32) R. R. Beal, "RCA developments in television," J. Soc. Mot. Pict. Eng., 
vol. 29, pp. 121-144; August, 1937. 

(33) Alfred N. Goldsmith, "Television and the motion picture theater," 
Intern. Proj., May, 1935. 

(34) O. H. Schade, "Electrooptical characteristics of television system," Part 
I, "Characteristics of vision and visual systems," RCA Rev., vol. 9, p. 5; March, 
1948; Part II, "Electrooptical specifications for television systems," RCA Rev., 
vol. 9, p. 245; June, 1948. 

Television from FUm 

(1) M. R. Boyer, "Test reel for television broadcast stations," J. Soc. Mot. 
Pict. Eng., vol. 49, pp. 391-395; November, 1947. 

(2) R. V. Little, "Film projectors for television," Intern. Proj., May, 1947. 

(3) Ralph V. Little, Jr., "Film projectors for television," J. Soc. Mot. Pict. 
Eng., vol. 48, pp. 93-111; February, 1947. 


(4) E. Meschter, "Television reproduction from negative films," J. Soc. Mot. 
Pict. Eng., vol. 47, pp. 165-182; August, 1946. 

(5) Ellsworth D. Cook, "General Electric television film projector," J. Soc. 
Mot. Pict. Eng., vol. 41, pp. 273-292; October, 1943. 

(6) R. B. Fuller and L. S. Rhodes, "Production of 16-mm motion pictures for 
television projection," J. Soc. Mot. Pict. Eng., vol. 39, pp. 195-202; September, 

(7) Axel G. Jensen, "Film scanner for use in television transmission tests," 
Proc. I.R.E., vol. 29, pp. 243-250; May, 1941. 

(8) Harry R. Lubcke, "Photographic aspects of television operations,"./. Soc. 
Mot. Pict. Eng., vol. 36, pp. 185-191; February, 1941. 

(9) C. Frederick Wolcott, "Problems in television image resolution," /. Soc. 
Mot. Pict. Eng., vol. 36, pp. 65-82; January, 1941. 

(10) R. L. Campbell, "Television control equipment for film transmission," 
/. Soc. Mot. Pict. Eng., vol. 33, pp. 677-690; December, 1939. 

(11) Peter C. Goldmark, "Continuous type television film scanner," /. Soc. 
Mot. Pict. Eng., vol. 33, pp. 18-26; July, 1939. 

(12) E. W. Engstrom, G. L. Beers, and A. V. Bedford, "Application of motion to television," J. Soc. Mot. Pict. Eng., vol. 33, pp. 3-18; July, 1939; 
RCA Rev., vol. 4, p. 48; July, 1939. 

(13) H. S. Bamford, "Non-intermittent projector for television film transmis- 
sion," /. Soc. Mot. Pict. Eng., vol. 31, pp. 453-462; November, 1938. 

(14) K. Thorn, "Neuer mechanischer Filmabtaster," Fernseh A. G. Hausmit- 
teilungen, vol. 1, p. 24; August, 1938. 

(15) H. E. Ives, "A multi-channel television apparatus," Bell Sys. Tech. Jour., 
vol. 10, p. 33; January, 1931. 

Film from Television 

(1) R. M. Fraser, "Motion picture photography of television images," RCA 
Rev., vol. 9, p. 202; June, 1948. 

(2) C. F. White and M. R. Boyer, "A new film for photographing the television 
monitor tube," J. Soc. Mot. Pict. Eng., vol. 47, pp. 152-165; August, 1946. 

(3) F. G. Albin, "Sensitometric aspect of television monitor-tube photogra- 
phy," /. Soc. Mot. Pict. Eng., vol. 51, pp. 595-613; December, 1948. 

(4) J. L. Boon, W. Feldman, and J. Stoiber, "Television recording camera," 
/. Soc. Mot. Pict. Eng., vol. 51, pp. 117-127; August, 1948. 

(5) Thomas T. Goldsmith, Jr., and Harry Milholland, "Television transcrip- 
tion by motion picture film," /. Soc. Mot. Pict. Eng., vol. 51, pp. 107-117; August, 

Color Television 

(1) W. H. Cherry, "Colorimetry in television," RCA Rev., vol. 8, pp. 427-460; 
September, 1947; J. Soc. Mot. Pict. Eng., vol. 51, pp. 613-643; December, 1948. 

(2) R. D. Kell, "An experimental simultaneous color-television system, Part 
I, Introduction," Proc. I.R.E., vol. 35, pp. 861-862; September, 1947. 

(3) G. C. Sziklai, R. C. Ballard, and A. C. Schroeder, "Part II, Pickup equip- 
ment," Proc. I.R.E., vol. 35, pp. 862-871; September, 1947. 


(4) K. R. Wendt, G. L. Fredendall, and A. C. Schroeder, "Part III, radio-fre- 
quency and reproducing equipment," Proc. I.R.E., vol. 35, pp. 871-875; Septem- 
ber, 1947. 

(5) Statements and Exhibits of CBS and RCA at FCC Hearing on Color Tele- 
vision, December 9, 1946. 

(6) UHF Television Systems, Reports by RMA Committees, Data Bureau, 
RMA, November 26, 1946. 

(7) Interim Report, UHF Color Television, RTPB Panel 6, RMA Television 
Systems Committee, Data Bureau, RMA, November 25, 1946. 

(8) "Simultaneous all electronic color television," RCA Rev., vol. 7, p. 459; 
December, 1946. 

(9) R. D. Kell, G. L. Fredendall, A. C. Schroeder, and R. C. Webb, "An ex- 
perimental color television system," RCA Rev., vol. 7, p. 141; June, 1946. 

(10) P. C. Goldmark, E. R. Piore, J. M. Hollywood, T. H. Chambers, and J. J. 
Reeves, "Color television Part II," Proc. I.R.E., vol. 3